linux/fs/xfs/xfs_vfsops.c

1391 lines
35 KiB
C
Raw Normal View History

/*
* Copyright (c) 2000-2005 Silicon Graphics, Inc.
* All Rights Reserved.
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation.
*
* This program is distributed in the hope that it would be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write the Free Software Foundation,
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "xfs.h"
#include "xfs_fs.h"
#include "xfs_types.h"
#include "xfs_bit.h"
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_btree.h"
#include "xfs_alloc.h"
#include "xfs_ialloc.h"
#include "xfs_quota.h"
#include "xfs_error.h"
#include "xfs_bmap.h"
#include "xfs_rw.h"
#include "xfs_refcache.h"
#include "xfs_buf_item.h"
#include "xfs_log_priv.h"
#include "xfs_dir2_trace.h"
#include "xfs_extfree_item.h"
#include "xfs_acl.h"
#include "xfs_attr.h"
#include "xfs_clnt.h"
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
#include "xfs_mru_cache.h"
#include "xfs_filestream.h"
#include "xfs_fsops.h"
#include "xfs_vnodeops.h"
#include "xfs_vfsops.h"
int __init
xfs_init(void)
{
#ifdef XFS_DABUF_DEBUG
extern spinlock_t xfs_dabuf_global_lock;
spin_lock_init(&xfs_dabuf_global_lock);
#endif
/*
* Initialize all of the zone allocators we use.
*/
xfs_bmap_free_item_zone = kmem_zone_init(sizeof(xfs_bmap_free_item_t),
"xfs_bmap_free_item");
xfs_btree_cur_zone = kmem_zone_init(sizeof(xfs_btree_cur_t),
"xfs_btree_cur");
xfs_trans_zone = kmem_zone_init(sizeof(xfs_trans_t), "xfs_trans");
xfs_da_state_zone =
kmem_zone_init(sizeof(xfs_da_state_t), "xfs_da_state");
xfs_dabuf_zone = kmem_zone_init(sizeof(xfs_dabuf_t), "xfs_dabuf");
xfs_ifork_zone = kmem_zone_init(sizeof(xfs_ifork_t), "xfs_ifork");
xfs_acl_zone_init(xfs_acl_zone, "xfs_acl");
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
xfs_mru_cache_init();
xfs_filestream_init();
/*
* The size of the zone allocated buf log item is the maximum
* size possible under XFS. This wastes a little bit of memory,
* but it is much faster.
*/
xfs_buf_item_zone =
kmem_zone_init((sizeof(xfs_buf_log_item_t) +
(((XFS_MAX_BLOCKSIZE / XFS_BLI_CHUNK) /
NBWORD) * sizeof(int))),
"xfs_buf_item");
xfs_efd_zone =
kmem_zone_init((sizeof(xfs_efd_log_item_t) +
((XFS_EFD_MAX_FAST_EXTENTS - 1) *
sizeof(xfs_extent_t))),
"xfs_efd_item");
xfs_efi_zone =
kmem_zone_init((sizeof(xfs_efi_log_item_t) +
((XFS_EFI_MAX_FAST_EXTENTS - 1) *
sizeof(xfs_extent_t))),
"xfs_efi_item");
/*
* These zones warrant special memory allocator hints
*/
xfs_inode_zone =
kmem_zone_init_flags(sizeof(xfs_inode_t), "xfs_inode",
KM_ZONE_HWALIGN | KM_ZONE_RECLAIM |
KM_ZONE_SPREAD, NULL);
xfs_ili_zone =
kmem_zone_init_flags(sizeof(xfs_inode_log_item_t), "xfs_ili",
KM_ZONE_SPREAD, NULL);
xfs_icluster_zone =
kmem_zone_init_flags(sizeof(xfs_icluster_t), "xfs_icluster",
KM_ZONE_SPREAD, NULL);
/*
* Allocate global trace buffers.
*/
#ifdef XFS_ALLOC_TRACE
xfs_alloc_trace_buf = ktrace_alloc(XFS_ALLOC_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_BMAP_TRACE
xfs_bmap_trace_buf = ktrace_alloc(XFS_BMAP_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_BMBT_TRACE
xfs_bmbt_trace_buf = ktrace_alloc(XFS_BMBT_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_ATTR_TRACE
xfs_attr_trace_buf = ktrace_alloc(XFS_ATTR_TRACE_SIZE, KM_SLEEP);
#endif
#ifdef XFS_DIR2_TRACE
xfs_dir2_trace_buf = ktrace_alloc(XFS_DIR2_GTRACE_SIZE, KM_SLEEP);
#endif
xfs_dir_startup();
#if (defined(DEBUG) || defined(INDUCE_IO_ERROR))
xfs_error_test_init();
#endif /* DEBUG || INDUCE_IO_ERROR */
xfs_init_procfs();
xfs_sysctl_register();
return 0;
}
void __exit
xfs_cleanup(void)
{
extern kmem_zone_t *xfs_inode_zone;
extern kmem_zone_t *xfs_efd_zone;
extern kmem_zone_t *xfs_efi_zone;
extern kmem_zone_t *xfs_icluster_zone;
xfs_cleanup_procfs();
xfs_sysctl_unregister();
xfs_refcache_destroy();
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
xfs_filestream_uninit();
xfs_mru_cache_uninit();
xfs_acl_zone_destroy(xfs_acl_zone);
#ifdef XFS_DIR2_TRACE
ktrace_free(xfs_dir2_trace_buf);
#endif
#ifdef XFS_ATTR_TRACE
ktrace_free(xfs_attr_trace_buf);
#endif
#ifdef XFS_BMBT_TRACE
ktrace_free(xfs_bmbt_trace_buf);
#endif
#ifdef XFS_BMAP_TRACE
ktrace_free(xfs_bmap_trace_buf);
#endif
#ifdef XFS_ALLOC_TRACE
ktrace_free(xfs_alloc_trace_buf);
#endif
kmem_zone_destroy(xfs_bmap_free_item_zone);
kmem_zone_destroy(xfs_btree_cur_zone);
kmem_zone_destroy(xfs_inode_zone);
kmem_zone_destroy(xfs_trans_zone);
kmem_zone_destroy(xfs_da_state_zone);
kmem_zone_destroy(xfs_dabuf_zone);
kmem_zone_destroy(xfs_buf_item_zone);
kmem_zone_destroy(xfs_efd_zone);
kmem_zone_destroy(xfs_efi_zone);
kmem_zone_destroy(xfs_ifork_zone);
kmem_zone_destroy(xfs_ili_zone);
kmem_zone_destroy(xfs_icluster_zone);
}
/*
* xfs_start_flags
*
* This function fills in xfs_mount_t fields based on mount args.
* Note: the superblock has _not_ yet been read in.
*/
STATIC int
xfs_start_flags(
struct xfs_mount_args *ap,
struct xfs_mount *mp)
{
/* Values are in BBs */
if ((ap->flags & XFSMNT_NOALIGN) != XFSMNT_NOALIGN) {
/*
* At this point the superblock has not been read
* in, therefore we do not know the block size.
* Before the mount call ends we will convert
* these to FSBs.
*/
mp->m_dalign = ap->sunit;
mp->m_swidth = ap->swidth;
}
if (ap->logbufs != -1 &&
ap->logbufs != 0 &&
(ap->logbufs < XLOG_MIN_ICLOGS ||
ap->logbufs > XLOG_MAX_ICLOGS)) {
cmn_err(CE_WARN,
"XFS: invalid logbufs value: %d [not %d-%d]",
ap->logbufs, XLOG_MIN_ICLOGS, XLOG_MAX_ICLOGS);
return XFS_ERROR(EINVAL);
}
mp->m_logbufs = ap->logbufs;
if (ap->logbufsize != -1 &&
ap->logbufsize != 0 &&
(ap->logbufsize < XLOG_MIN_RECORD_BSIZE ||
ap->logbufsize > XLOG_MAX_RECORD_BSIZE ||
!is_power_of_2(ap->logbufsize))) {
cmn_err(CE_WARN,
"XFS: invalid logbufsize: %d [not 16k,32k,64k,128k or 256k]",
ap->logbufsize);
return XFS_ERROR(EINVAL);
}
mp->m_logbsize = ap->logbufsize;
mp->m_fsname_len = strlen(ap->fsname) + 1;
mp->m_fsname = kmem_alloc(mp->m_fsname_len, KM_SLEEP);
strcpy(mp->m_fsname, ap->fsname);
if (ap->rtname[0]) {
mp->m_rtname = kmem_alloc(strlen(ap->rtname) + 1, KM_SLEEP);
strcpy(mp->m_rtname, ap->rtname);
}
if (ap->logname[0]) {
mp->m_logname = kmem_alloc(strlen(ap->logname) + 1, KM_SLEEP);
strcpy(mp->m_logname, ap->logname);
}
if (ap->flags & XFSMNT_WSYNC)
mp->m_flags |= XFS_MOUNT_WSYNC;
#if XFS_BIG_INUMS
if (ap->flags & XFSMNT_INO64) {
mp->m_flags |= XFS_MOUNT_INO64;
mp->m_inoadd = XFS_INO64_OFFSET;
}
#endif
if (ap->flags & XFSMNT_RETERR)
mp->m_flags |= XFS_MOUNT_RETERR;
if (ap->flags & XFSMNT_NOALIGN)
mp->m_flags |= XFS_MOUNT_NOALIGN;
if (ap->flags & XFSMNT_SWALLOC)
mp->m_flags |= XFS_MOUNT_SWALLOC;
if (ap->flags & XFSMNT_OSYNCISOSYNC)
mp->m_flags |= XFS_MOUNT_OSYNCISOSYNC;
if (ap->flags & XFSMNT_32BITINODES)
mp->m_flags |= XFS_MOUNT_32BITINODES;
if (ap->flags & XFSMNT_IOSIZE) {
if (ap->iosizelog > XFS_MAX_IO_LOG ||
ap->iosizelog < XFS_MIN_IO_LOG) {
cmn_err(CE_WARN,
"XFS: invalid log iosize: %d [not %d-%d]",
ap->iosizelog, XFS_MIN_IO_LOG,
XFS_MAX_IO_LOG);
return XFS_ERROR(EINVAL);
}
mp->m_flags |= XFS_MOUNT_DFLT_IOSIZE;
mp->m_readio_log = mp->m_writeio_log = ap->iosizelog;
}
if (ap->flags & XFSMNT_IDELETE)
mp->m_flags |= XFS_MOUNT_IDELETE;
if (ap->flags & XFSMNT_DIRSYNC)
mp->m_flags |= XFS_MOUNT_DIRSYNC;
if (ap->flags & XFSMNT_ATTR2)
mp->m_flags |= XFS_MOUNT_ATTR2;
if (ap->flags2 & XFSMNT2_COMPAT_IOSIZE)
mp->m_flags |= XFS_MOUNT_COMPAT_IOSIZE;
/*
* no recovery flag requires a read-only mount
*/
if (ap->flags & XFSMNT_NORECOVERY) {
if (!(mp->m_flags & XFS_MOUNT_RDONLY)) {
cmn_err(CE_WARN,
"XFS: tried to mount a FS read-write without recovery!");
return XFS_ERROR(EINVAL);
}
mp->m_flags |= XFS_MOUNT_NORECOVERY;
}
if (ap->flags & XFSMNT_NOUUID)
mp->m_flags |= XFS_MOUNT_NOUUID;
if (ap->flags & XFSMNT_BARRIER)
mp->m_flags |= XFS_MOUNT_BARRIER;
else
mp->m_flags &= ~XFS_MOUNT_BARRIER;
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
if (ap->flags2 & XFSMNT2_FILESTREAMS)
mp->m_flags |= XFS_MOUNT_FILESTREAMS;
if (ap->flags & XFSMNT_DMAPI)
mp->m_flags |= XFS_MOUNT_DMAPI;
return 0;
}
/*
* This function fills in xfs_mount_t fields based on mount args.
* Note: the superblock _has_ now been read in.
*/
STATIC int
xfs_finish_flags(
struct xfs_mount_args *ap,
struct xfs_mount *mp)
{
int ronly = (mp->m_flags & XFS_MOUNT_RDONLY);
/* Fail a mount where the logbuf is smaller then the log stripe */
if (XFS_SB_VERSION_HASLOGV2(&mp->m_sb)) {
if ((ap->logbufsize <= 0) &&
(mp->m_sb.sb_logsunit > XLOG_BIG_RECORD_BSIZE)) {
mp->m_logbsize = mp->m_sb.sb_logsunit;
} else if (ap->logbufsize > 0 &&
ap->logbufsize < mp->m_sb.sb_logsunit) {
cmn_err(CE_WARN,
"XFS: logbuf size must be greater than or equal to log stripe size");
return XFS_ERROR(EINVAL);
}
} else {
/* Fail a mount if the logbuf is larger than 32K */
if (ap->logbufsize > XLOG_BIG_RECORD_BSIZE) {
cmn_err(CE_WARN,
"XFS: logbuf size for version 1 logs must be 16K or 32K");
return XFS_ERROR(EINVAL);
}
}
if (XFS_SB_VERSION_HASATTR2(&mp->m_sb)) {
mp->m_flags |= XFS_MOUNT_ATTR2;
}
/*
* prohibit r/w mounts of read-only filesystems
*/
if ((mp->m_sb.sb_flags & XFS_SBF_READONLY) && !ronly) {
cmn_err(CE_WARN,
"XFS: cannot mount a read-only filesystem as read-write");
return XFS_ERROR(EROFS);
}
/*
* check for shared mount.
*/
if (ap->flags & XFSMNT_SHARED) {
if (!XFS_SB_VERSION_HASSHARED(&mp->m_sb))
return XFS_ERROR(EINVAL);
/*
* For IRIX 6.5, shared mounts must have the shared
* version bit set, have the persistent readonly
* field set, must be version 0 and can only be mounted
* read-only.
*/
if (!ronly || !(mp->m_sb.sb_flags & XFS_SBF_READONLY) ||
(mp->m_sb.sb_shared_vn != 0))
return XFS_ERROR(EINVAL);
mp->m_flags |= XFS_MOUNT_SHARED;
/*
* Shared XFS V0 can't deal with DMI. Return EINVAL.
*/
if (mp->m_sb.sb_shared_vn == 0 && (ap->flags & XFSMNT_DMAPI))
return XFS_ERROR(EINVAL);
}
if (ap->flags & XFSMNT_UQUOTA) {
mp->m_qflags |= (XFS_UQUOTA_ACCT | XFS_UQUOTA_ACTIVE);
if (ap->flags & XFSMNT_UQUOTAENF)
mp->m_qflags |= XFS_UQUOTA_ENFD;
}
if (ap->flags & XFSMNT_GQUOTA) {
mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
if (ap->flags & XFSMNT_GQUOTAENF)
mp->m_qflags |= XFS_OQUOTA_ENFD;
} else if (ap->flags & XFSMNT_PQUOTA) {
mp->m_qflags |= (XFS_PQUOTA_ACCT | XFS_PQUOTA_ACTIVE);
if (ap->flags & XFSMNT_PQUOTAENF)
mp->m_qflags |= XFS_OQUOTA_ENFD;
}
return 0;
}
/*
* xfs_mount
*
* The file system configurations are:
* (1) device (partition) with data and internal log
* (2) logical volume with data and log subvolumes.
* (3) logical volume with data, log, and realtime subvolumes.
*
* We only have to handle opening the log and realtime volumes here if
* they are present. The data subvolume has already been opened by
* get_sb_bdev() and is stored in vfsp->vfs_super->s_bdev.
*/
int
xfs_mount(
struct xfs_mount *mp,
struct xfs_mount_args *args,
cred_t *credp)
{
struct block_device *ddev, *logdev, *rtdev;
int flags = 0, error;
ddev = mp->m_super->s_bdev;
logdev = rtdev = NULL;
error = xfs_dmops_get(mp, args);
if (error)
return error;
error = xfs_qmops_get(mp, args);
if (error)
return error;
if (args->flags & XFSMNT_QUIET)
flags |= XFS_MFSI_QUIET;
/*
* Open real time and log devices - order is important.
*/
if (args->logname[0]) {
error = xfs_blkdev_get(mp, args->logname, &logdev);
if (error)
return error;
}
if (args->rtname[0]) {
error = xfs_blkdev_get(mp, args->rtname, &rtdev);
if (error) {
xfs_blkdev_put(logdev);
return error;
}
if (rtdev == ddev || rtdev == logdev) {
cmn_err(CE_WARN,
"XFS: Cannot mount filesystem with identical rtdev and ddev/logdev.");
xfs_blkdev_put(logdev);
xfs_blkdev_put(rtdev);
return EINVAL;
}
}
/*
* Setup xfs_mount buffer target pointers
*/
error = ENOMEM;
mp->m_ddev_targp = xfs_alloc_buftarg(ddev, 0);
if (!mp->m_ddev_targp) {
xfs_blkdev_put(logdev);
xfs_blkdev_put(rtdev);
return error;
}
if (rtdev) {
mp->m_rtdev_targp = xfs_alloc_buftarg(rtdev, 1);
if (!mp->m_rtdev_targp) {
xfs_blkdev_put(logdev);
xfs_blkdev_put(rtdev);
goto error0;
}
}
mp->m_logdev_targp = (logdev && logdev != ddev) ?
xfs_alloc_buftarg(logdev, 1) : mp->m_ddev_targp;
if (!mp->m_logdev_targp) {
xfs_blkdev_put(logdev);
xfs_blkdev_put(rtdev);
goto error0;
}
/*
* Setup flags based on mount(2) options and then the superblock
*/
error = xfs_start_flags(args, mp);
if (error)
goto error1;
error = xfs_readsb(mp, flags);
if (error)
goto error1;
error = xfs_finish_flags(args, mp);
if (error)
goto error2;
/*
* Setup xfs_mount buffer target pointers based on superblock
*/
error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_blocksize,
mp->m_sb.sb_sectsize);
if (!error && logdev && logdev != ddev) {
unsigned int log_sector_size = BBSIZE;
if (XFS_SB_VERSION_HASSECTOR(&mp->m_sb))
log_sector_size = mp->m_sb.sb_logsectsize;
error = xfs_setsize_buftarg(mp->m_logdev_targp,
mp->m_sb.sb_blocksize,
log_sector_size);
}
if (!error && rtdev)
error = xfs_setsize_buftarg(mp->m_rtdev_targp,
mp->m_sb.sb_blocksize,
mp->m_sb.sb_sectsize);
if (error)
goto error2;
if (mp->m_flags & XFS_MOUNT_BARRIER)
xfs_mountfs_check_barriers(mp);
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
if ((error = xfs_filestream_mount(mp)))
goto error2;
error = xfs_mountfs(mp, flags);
if (error)
goto error2;
XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, args->mtpt, args->fsname);
return 0;
error2:
if (mp->m_sb_bp)
xfs_freesb(mp);
error1:
xfs_binval(mp->m_ddev_targp);
if (logdev && logdev != ddev)
xfs_binval(mp->m_logdev_targp);
if (rtdev)
xfs_binval(mp->m_rtdev_targp);
error0:
xfs_unmountfs_close(mp, credp);
xfs_qmops_put(mp);
xfs_dmops_put(mp);
return error;
}
int
xfs_unmount(
xfs_mount_t *mp,
int flags,
cred_t *credp)
{
xfs_inode_t *rip;
bhv_vnode_t *rvp;
int unmount_event_wanted = 0;
int unmount_event_flags = 0;
int xfs_unmountfs_needed = 0;
int error;
rip = mp->m_rootip;
rvp = XFS_ITOV(rip);
#ifdef HAVE_DMAPI
if (mp->m_flags & XFS_MOUNT_DMAPI) {
error = XFS_SEND_PREUNMOUNT(mp,
rvp, DM_RIGHT_NULL, rvp, DM_RIGHT_NULL,
NULL, NULL, 0, 0,
(mp->m_dmevmask & (1<<DM_EVENT_PREUNMOUNT))?
0:DM_FLAGS_UNWANTED);
if (error)
return XFS_ERROR(error);
unmount_event_wanted = 1;
unmount_event_flags = (mp->m_dmevmask & (1<<DM_EVENT_UNMOUNT))?
0 : DM_FLAGS_UNWANTED;
}
#endif
/*
* First blow any referenced inode from this file system
* out of the reference cache, and delete the timer.
*/
xfs_refcache_purge_mp(mp);
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
/*
* Blow away any referenced inode in the filestreams cache.
* This can and will cause log traffic as inodes go inactive
* here.
*/
xfs_filestream_unmount(mp);
XFS_bflush(mp->m_ddev_targp);
error = xfs_unmount_flush(mp, 0);
if (error)
goto out;
ASSERT(vn_count(rvp) == 1);
/*
* Drop the reference count
*/
VN_RELE(rvp);
/*
* If we're forcing a shutdown, typically because of a media error,
* we want to make sure we invalidate dirty pages that belong to
* referenced vnodes as well.
*/
if (XFS_FORCED_SHUTDOWN(mp)) {
error = xfs_sync(mp, SYNC_WAIT | SYNC_CLOSE);
ASSERT(error != EFSCORRUPTED);
}
xfs_unmountfs_needed = 1;
out:
/* Send DMAPI event, if required.
* Then do xfs_unmountfs() if needed.
* Then return error (or zero).
*/
if (unmount_event_wanted) {
/* Note: mp structure must still exist for
* XFS_SEND_UNMOUNT() call.
*/
XFS_SEND_UNMOUNT(mp, error == 0 ? rvp : NULL,
DM_RIGHT_NULL, 0, error, unmount_event_flags);
}
if (xfs_unmountfs_needed) {
/*
* Call common unmount function to flush to disk
* and free the super block buffer & mount structures.
*/
xfs_unmountfs(mp, credp);
xfs_qmops_put(mp);
xfs_dmops_put(mp);
kmem_free(mp, sizeof(xfs_mount_t));
}
return XFS_ERROR(error);
}
STATIC int
xfs_quiesce_fs(
xfs_mount_t *mp)
{
int count = 0, pincount;
xfs_refcache_purge_mp(mp);
xfs_flush_buftarg(mp->m_ddev_targp, 0);
xfs_finish_reclaim_all(mp, 0);
/* This loop must run at least twice.
* The first instance of the loop will flush
* most meta data but that will generate more
* meta data (typically directory updates).
* Which then must be flushed and logged before
* we can write the unmount record.
*/
do {
xfs_syncsub(mp, SYNC_INODE_QUIESCE, NULL);
pincount = xfs_flush_buftarg(mp->m_ddev_targp, 1);
if (!pincount) {
delay(50);
count++;
}
} while (count < 2);
return 0;
}
/*
* Second stage of a quiesce. The data is already synced, now we have to take
* care of the metadata. New transactions are already blocked, so we need to
* wait for any remaining transactions to drain out before proceding.
*/
void
xfs_attr_quiesce(
xfs_mount_t *mp)
{
/* wait for all modifications to complete */
while (atomic_read(&mp->m_active_trans) > 0)
delay(100);
/* flush inodes and push all remaining buffers out to disk */
xfs_quiesce_fs(mp);
ASSERT_ALWAYS(atomic_read(&mp->m_active_trans) == 0);
/* Push the superblock and write an unmount record */
xfs_log_sbcount(mp, 1);
xfs_log_unmount_write(mp);
xfs_unmountfs_writesb(mp);
}
int
xfs_mntupdate(
struct xfs_mount *mp,
int *flags,
struct xfs_mount_args *args)
{
if (!(*flags & MS_RDONLY)) { /* rw/ro -> rw */
if (mp->m_flags & XFS_MOUNT_RDONLY)
mp->m_flags &= ~XFS_MOUNT_RDONLY;
if (args->flags & XFSMNT_BARRIER) {
mp->m_flags |= XFS_MOUNT_BARRIER;
xfs_mountfs_check_barriers(mp);
} else {
mp->m_flags &= ~XFS_MOUNT_BARRIER;
}
} else if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { /* rw -> ro */
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
xfs_filestream_flush(mp);
xfs_sync(mp, SYNC_DATA_QUIESCE);
xfs_attr_quiesce(mp);
mp->m_flags |= XFS_MOUNT_RDONLY;
}
return 0;
}
/*
* xfs_unmount_flush implements a set of flush operation on special
* inodes, which are needed as a separate set of operations so that
* they can be called as part of relocation process.
*/
int
xfs_unmount_flush(
xfs_mount_t *mp, /* Mount structure we are getting
rid of. */
int relocation) /* Called from vfs relocation. */
{
xfs_inode_t *rip = mp->m_rootip;
xfs_inode_t *rbmip;
xfs_inode_t *rsumip = NULL;
bhv_vnode_t *rvp = XFS_ITOV(rip);
int error;
xfs_ilock(rip, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
xfs_iflock(rip);
/*
* Flush out the real time inodes.
*/
if ((rbmip = mp->m_rbmip) != NULL) {
xfs_ilock(rbmip, XFS_ILOCK_EXCL);
xfs_iflock(rbmip);
error = xfs_iflush(rbmip, XFS_IFLUSH_SYNC);
xfs_iunlock(rbmip, XFS_ILOCK_EXCL);
if (error == EFSCORRUPTED)
goto fscorrupt_out;
ASSERT(vn_count(XFS_ITOV(rbmip)) == 1);
rsumip = mp->m_rsumip;
xfs_ilock(rsumip, XFS_ILOCK_EXCL);
xfs_iflock(rsumip);
error = xfs_iflush(rsumip, XFS_IFLUSH_SYNC);
xfs_iunlock(rsumip, XFS_ILOCK_EXCL);
if (error == EFSCORRUPTED)
goto fscorrupt_out;
ASSERT(vn_count(XFS_ITOV(rsumip)) == 1);
}
/*
* Synchronously flush root inode to disk
*/
error = xfs_iflush(rip, XFS_IFLUSH_SYNC);
if (error == EFSCORRUPTED)
goto fscorrupt_out2;
if (vn_count(rvp) != 1 && !relocation) {
xfs_iunlock(rip, XFS_ILOCK_EXCL);
return XFS_ERROR(EBUSY);
}
/*
* Release dquot that rootinode, rbmino and rsumino might be holding,
* flush and purge the quota inodes.
*/
error = XFS_QM_UNMOUNT(mp);
if (error == EFSCORRUPTED)
goto fscorrupt_out2;
if (rbmip) {
VN_RELE(XFS_ITOV(rbmip));
VN_RELE(XFS_ITOV(rsumip));
}
xfs_iunlock(rip, XFS_ILOCK_EXCL);
return 0;
fscorrupt_out:
xfs_ifunlock(rip);
fscorrupt_out2:
xfs_iunlock(rip, XFS_ILOCK_EXCL);
return XFS_ERROR(EFSCORRUPTED);
}
/*
* xfs_sync flushes any pending I/O to file system vfsp.
*
* This routine is called by vfs_sync() to make sure that things make it
* out to disk eventually, on sync() system calls to flush out everything,
* and when the file system is unmounted. For the vfs_sync() case, all
* we really need to do is sync out the log to make all of our meta-data
* updates permanent (except for timestamps). For calls from pflushd(),
* dirty pages are kept moving by calling pdflush() on the inodes
* containing them. We also flush the inodes that we can lock without
* sleeping and the superblock if we can lock it without sleeping from
* vfs_sync() so that items at the tail of the log are always moving out.
*
* Flags:
* SYNC_BDFLUSH - We're being called from vfs_sync() so we don't want
* to sleep if we can help it. All we really need
* to do is ensure that the log is synced at least
* periodically. We also push the inodes and
* superblock if we can lock them without sleeping
* and they are not pinned.
* SYNC_ATTR - We need to flush the inodes. If SYNC_BDFLUSH is not
* set, then we really want to lock each inode and flush
* it.
* SYNC_WAIT - All the flushes that take place in this call should
* be synchronous.
* SYNC_DELWRI - This tells us to push dirty pages associated with
* inodes. SYNC_WAIT and SYNC_BDFLUSH are used to
* determine if they should be flushed sync, async, or
* delwri.
* SYNC_CLOSE - This flag is passed when the system is being
* unmounted. We should sync and invalidate everything.
* SYNC_FSDATA - This indicates that the caller would like to make
* sure the superblock is safe on disk. We can ensure
* this by simply making sure the log gets flushed
* if SYNC_BDFLUSH is set, and by actually writing it
* out otherwise.
* SYNC_IOWAIT - The caller wants us to wait for all data I/O to complete
* before we return (including direct I/O). Forms the drain
* side of the write barrier needed to safely quiesce the
* filesystem.
*
*/
int
xfs_sync(
xfs_mount_t *mp,
int flags)
{
int error;
/*
* Get the Quota Manager to flush the dquots.
*
* If XFS quota support is not enabled or this filesystem
* instance does not use quotas XFS_QM_DQSYNC will always
* return zero.
*/
error = XFS_QM_DQSYNC(mp, flags);
if (error) {
/*
* If we got an IO error, we will be shutting down.
* So, there's nothing more for us to do here.
*/
ASSERT(error != EIO || XFS_FORCED_SHUTDOWN(mp));
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(error);
}
[XFS] Concurrent Multi-File Data Streams In media spaces, video is often stored in a frame-per-file format. When dealing with uncompressed realtime HD video streams in this format, it is crucial that files do not get fragmented and that multiple files a placed contiguously on disk. When multiple streams are being ingested and played out at the same time, it is critical that the filesystem does not cross the streams and interleave them together as this creates seek and readahead cache miss latency and prevents both ingest and playout from meeting frame rate targets. This patch set creates a "stream of files" concept into the allocator to place all the data from a single stream contiguously on disk so that RAID array readahead can be used effectively. Each additional stream gets placed in different allocation groups within the filesystem, thereby ensuring that we don't cross any streams. When an AG fills up, we select a new AG for the stream that is not in use. The core of the functionality is the stream tracking - each inode that we create in a directory needs to be associated with the directories' stream. Hence every time we create a file, we look up the directories' stream object and associate the new file with that object. Once we have a stream object for a file, we use the AG that the stream object point to for allocations. If we can't allocate in that AG (e.g. it is full) we move the entire stream to another AG. Other inodes in the same stream are moved to the new AG on their next allocation (i.e. lazy update). Stream objects are kept in a cache and hold a reference on the inode. Hence the inode cannot be reclaimed while there is an outstanding stream reference. This means that on unlink we need to remove the stream association and we also need to flush all the associations on certain events that want to reclaim all unreferenced inodes (e.g. filesystem freeze). SGI-PV: 964469 SGI-Modid: xfs-linux-melb:xfs-kern:29096a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Barry Naujok <bnaujok@sgi.com> Signed-off-by: Donald Douwsma <donaldd@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com> Signed-off-by: Vlad Apostolov <vapo@sgi.com>
2007-07-11 09:09:12 +08:00
if (flags & SYNC_IOWAIT)
xfs_filestream_flush(mp);
return xfs_syncsub(mp, flags, NULL);
}
/*
* xfs sync routine for internal use
*
* This routine supports all of the flags defined for the generic vfs_sync
* interface as explained above under xfs_sync.
*
*/
int
xfs_sync_inodes(
xfs_mount_t *mp,
int flags,
int *bypassed)
{
xfs_inode_t *ip = NULL;
bhv_vnode_t *vp = NULL;
int error;
int last_error;
uint64_t fflag;
uint lock_flags;
uint base_lock_flags;
boolean_t mount_locked;
boolean_t vnode_refed;
int preempt;
xfs_iptr_t *ipointer;
#ifdef DEBUG
boolean_t ipointer_in = B_FALSE;
#define IPOINTER_SET ipointer_in = B_TRUE
#define IPOINTER_CLR ipointer_in = B_FALSE
#else
#define IPOINTER_SET
#define IPOINTER_CLR
#endif
/* Insert a marker record into the inode list after inode ip. The list
* must be locked when this is called. After the call the list will no
* longer be locked.
*/
#define IPOINTER_INSERT(ip, mp) { \
ASSERT(ipointer_in == B_FALSE); \
ipointer->ip_mnext = ip->i_mnext; \
ipointer->ip_mprev = ip; \
ip->i_mnext = (xfs_inode_t *)ipointer; \
ipointer->ip_mnext->i_mprev = (xfs_inode_t *)ipointer; \
preempt = 0; \
XFS_MOUNT_IUNLOCK(mp); \
mount_locked = B_FALSE; \
IPOINTER_SET; \
}
/* Remove the marker from the inode list. If the marker was the only item
* in the list then there are no remaining inodes and we should zero out
* the whole list. If we are the current head of the list then move the head
* past us.
*/
#define IPOINTER_REMOVE(ip, mp) { \
ASSERT(ipointer_in == B_TRUE); \
if (ipointer->ip_mnext != (xfs_inode_t *)ipointer) { \
ip = ipointer->ip_mnext; \
ip->i_mprev = ipointer->ip_mprev; \
ipointer->ip_mprev->i_mnext = ip; \
if (mp->m_inodes == (xfs_inode_t *)ipointer) { \
mp->m_inodes = ip; \
} \
} else { \
ASSERT(mp->m_inodes == (xfs_inode_t *)ipointer); \
mp->m_inodes = NULL; \
ip = NULL; \
} \
IPOINTER_CLR; \
}
#define XFS_PREEMPT_MASK 0x7f
ASSERT(!(flags & SYNC_BDFLUSH));
if (bypassed)
*bypassed = 0;
if (mp->m_flags & XFS_MOUNT_RDONLY)
return 0;
error = 0;
last_error = 0;
preempt = 0;
/* Allocate a reference marker */
ipointer = (xfs_iptr_t *)kmem_zalloc(sizeof(xfs_iptr_t), KM_SLEEP);
fflag = XFS_B_ASYNC; /* default is don't wait */
if (flags & SYNC_DELWRI)
fflag = XFS_B_DELWRI;
if (flags & SYNC_WAIT)
fflag = 0; /* synchronous overrides all */
base_lock_flags = XFS_ILOCK_SHARED;
if (flags & (SYNC_DELWRI | SYNC_CLOSE)) {
/*
* We need the I/O lock if we're going to call any of
* the flush/inval routines.
*/
base_lock_flags |= XFS_IOLOCK_SHARED;
}
XFS_MOUNT_ILOCK(mp);
ip = mp->m_inodes;
mount_locked = B_TRUE;
vnode_refed = B_FALSE;
IPOINTER_CLR;
do {
ASSERT(ipointer_in == B_FALSE);
ASSERT(vnode_refed == B_FALSE);
lock_flags = base_lock_flags;
/*
* There were no inodes in the list, just break out
* of the loop.
*/
if (ip == NULL) {
break;
}
/*
* We found another sync thread marker - skip it
*/
if (ip->i_mount == NULL) {
ip = ip->i_mnext;
continue;
}
vp = XFS_ITOV_NULL(ip);
/*
* If the vnode is gone then this is being torn down,
* call reclaim if it is flushed, else let regular flush
* code deal with it later in the loop.
*/
if (vp == NULL) {
/* Skip ones already in reclaim */
if (ip->i_flags & XFS_IRECLAIM) {
ip = ip->i_mnext;
continue;
}
if (xfs_ilock_nowait(ip, XFS_ILOCK_EXCL) == 0) {
ip = ip->i_mnext;
} else if ((xfs_ipincount(ip) == 0) &&
xfs_iflock_nowait(ip)) {
IPOINTER_INSERT(ip, mp);
xfs_finish_reclaim(ip, 1,
XFS_IFLUSH_DELWRI_ELSE_ASYNC);
XFS_MOUNT_ILOCK(mp);
mount_locked = B_TRUE;
IPOINTER_REMOVE(ip, mp);
} else {
xfs_iunlock(ip, XFS_ILOCK_EXCL);
ip = ip->i_mnext;
}
continue;
}
if (VN_BAD(vp)) {
ip = ip->i_mnext;
continue;
}
if (XFS_FORCED_SHUTDOWN(mp) && !(flags & SYNC_CLOSE)) {
XFS_MOUNT_IUNLOCK(mp);
kmem_free(ipointer, sizeof(xfs_iptr_t));
return 0;
}
/*
* Try to lock without sleeping. We're out of order with
* the inode list lock here, so if we fail we need to drop
* the mount lock and try again. If we're called from
* bdflush() here, then don't bother.
*
* The inode lock here actually coordinates with the
* almost spurious inode lock in xfs_ireclaim() to prevent
* the vnode we handle here without a reference from
* being freed while we reference it. If we lock the inode
* while it's on the mount list here, then the spurious inode
* lock in xfs_ireclaim() after the inode is pulled from
* the mount list will sleep until we release it here.
* This keeps the vnode from being freed while we reference
* it.
*/
if (xfs_ilock_nowait(ip, lock_flags) == 0) {
if (vp == NULL) {
ip = ip->i_mnext;
continue;
}
vp = vn_grab(vp);
if (vp == NULL) {
ip = ip->i_mnext;
continue;
}
IPOINTER_INSERT(ip, mp);
xfs_ilock(ip, lock_flags);
ASSERT(vp == XFS_ITOV(ip));
ASSERT(ip->i_mount == mp);
vnode_refed = B_TRUE;
}
/* From here on in the loop we may have a marker record
* in the inode list.
*/
/*
* If we have to flush data or wait for I/O completion
* we need to drop the ilock that we currently hold.
* If we need to drop the lock, insert a marker if we
* have not already done so.
*/
if ((flags & (SYNC_CLOSE|SYNC_IOWAIT)) ||
((flags & SYNC_DELWRI) && VN_DIRTY(vp))) {
if (mount_locked) {
IPOINTER_INSERT(ip, mp);
}
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (flags & SYNC_CLOSE) {
/* Shutdown case. Flush and invalidate. */
if (XFS_FORCED_SHUTDOWN(mp))
xfs_tosspages(ip, 0, -1,
FI_REMAPF);
else
error = xfs_flushinval_pages(ip,
0, -1, FI_REMAPF);
} else if ((flags & SYNC_DELWRI) && VN_DIRTY(vp)) {
error = xfs_flush_pages(ip, 0,
-1, fflag, FI_NONE);
}
/*
* When freezing, we need to wait ensure all I/O (including direct
* I/O) is complete to ensure no further data modification can take
* place after this point
*/
if (flags & SYNC_IOWAIT)
vn_iowait(ip);
xfs_ilock(ip, XFS_ILOCK_SHARED);
}
if ((flags & SYNC_ATTR) &&
(ip->i_update_core ||
(ip->i_itemp && ip->i_itemp->ili_format.ilf_fields))) {
if (mount_locked)
IPOINTER_INSERT(ip, mp);
if (flags & SYNC_WAIT) {
xfs_iflock(ip);
error = xfs_iflush(ip, XFS_IFLUSH_SYNC);
/*
* If we can't acquire the flush lock, then the inode
* is already being flushed so don't bother waiting.
*
* If we can lock it then do a delwri flush so we can
* combine multiple inode flushes in each disk write.
*/
} else if (xfs_iflock_nowait(ip)) {
error = xfs_iflush(ip, XFS_IFLUSH_DELWRI);
} else if (bypassed) {
(*bypassed)++;
}
}
if (lock_flags != 0) {
xfs_iunlock(ip, lock_flags);
}
if (vnode_refed) {
/*
* If we had to take a reference on the vnode
* above, then wait until after we've unlocked
* the inode to release the reference. This is
* because we can be already holding the inode
* lock when VN_RELE() calls xfs_inactive().
*
* Make sure to drop the mount lock before calling
* VN_RELE() so that we don't trip over ourselves if
* we have to go for the mount lock again in the
* inactive code.
*/
if (mount_locked) {
IPOINTER_INSERT(ip, mp);
}
VN_RELE(vp);
vnode_refed = B_FALSE;
}
if (error) {
last_error = error;
}
/*
* bail out if the filesystem is corrupted.
*/
if (error == EFSCORRUPTED) {
if (!mount_locked) {
XFS_MOUNT_ILOCK(mp);
IPOINTER_REMOVE(ip, mp);
}
XFS_MOUNT_IUNLOCK(mp);
ASSERT(ipointer_in == B_FALSE);
kmem_free(ipointer, sizeof(xfs_iptr_t));
return XFS_ERROR(error);
}
/* Let other threads have a chance at the mount lock
* if we have looped many times without dropping the
* lock.
*/
if ((++preempt & XFS_PREEMPT_MASK) == 0) {
if (mount_locked) {
IPOINTER_INSERT(ip, mp);
}
}
if (mount_locked == B_FALSE) {
XFS_MOUNT_ILOCK(mp);
mount_locked = B_TRUE;
IPOINTER_REMOVE(ip, mp);
continue;
}
ASSERT(ipointer_in == B_FALSE);
ip = ip->i_mnext;
} while (ip != mp->m_inodes);
XFS_MOUNT_IUNLOCK(mp);
ASSERT(ipointer_in == B_FALSE);
kmem_free(ipointer, sizeof(xfs_iptr_t));
return XFS_ERROR(last_error);
}
/*
* xfs sync routine for internal use
*
* This routine supports all of the flags defined for the generic vfs_sync
* interface as explained above under xfs_sync.
*
*/
int
xfs_syncsub(
xfs_mount_t *mp,
int flags,
int *bypassed)
{
int error = 0;
int last_error = 0;
uint log_flags = XFS_LOG_FORCE;
xfs_buf_t *bp;
xfs_buf_log_item_t *bip;
/*
* Sync out the log. This ensures that the log is periodically
* flushed even if there is not enough activity to fill it up.
*/
if (flags & SYNC_WAIT)
log_flags |= XFS_LOG_SYNC;
xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
if (flags & (SYNC_ATTR|SYNC_DELWRI)) {
if (flags & SYNC_BDFLUSH)
xfs_finish_reclaim_all(mp, 1);
else
error = xfs_sync_inodes(mp, flags, bypassed);
}
/*
* Flushing out dirty data above probably generated more
* log activity, so if this isn't vfs_sync() then flush
* the log again.
*/
if (flags & SYNC_DELWRI) {
xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
}
if (flags & SYNC_FSDATA) {
/*
* If this is vfs_sync() then only sync the superblock
* if we can lock it without sleeping and it is not pinned.
*/
if (flags & SYNC_BDFLUSH) {
bp = xfs_getsb(mp, XFS_BUF_TRYLOCK);
if (bp != NULL) {
bip = XFS_BUF_FSPRIVATE(bp,xfs_buf_log_item_t*);
if ((bip != NULL) &&
xfs_buf_item_dirty(bip)) {
if (!(XFS_BUF_ISPINNED(bp))) {
XFS_BUF_ASYNC(bp);
error = xfs_bwrite(mp, bp);
} else {
xfs_buf_relse(bp);
}
} else {
xfs_buf_relse(bp);
}
}
} else {
bp = xfs_getsb(mp, 0);
/*
* If the buffer is pinned then push on the log so
* we won't get stuck waiting in the write for
* someone, maybe ourselves, to flush the log.
* Even though we just pushed the log above, we
* did not have the superblock buffer locked at
* that point so it can become pinned in between
* there and here.
*/
if (XFS_BUF_ISPINNED(bp))
xfs_log_force(mp, (xfs_lsn_t)0, XFS_LOG_FORCE);
if (flags & SYNC_WAIT)
XFS_BUF_UNASYNC(bp);
else
XFS_BUF_ASYNC(bp);
error = xfs_bwrite(mp, bp);
}
if (error) {
last_error = error;
}
}
/*
* If this is the periodic sync, then kick some entries out of
* the reference cache. This ensures that idle entries are
* eventually kicked out of the cache.
*/
if (flags & SYNC_REFCACHE) {
if (flags & SYNC_WAIT)
xfs_refcache_purge_mp(mp);
else
xfs_refcache_purge_some(mp);
}
[XFS] Lazy Superblock Counters When we have a couple of hundred transactions on the fly at once, they all typically modify the on disk superblock in some way. create/unclink/mkdir/rmdir modify inode counts, allocation/freeing modify free block counts. When these counts are modified in a transaction, they must eventually lock the superblock buffer and apply the mods. The buffer then remains locked until the transaction is committed into the incore log buffer. The result of this is that with enough transactions on the fly the incore superblock buffer becomes a bottleneck. The result of contention on the incore superblock buffer is that transaction rates fall - the more pressure that is put on the superblock buffer, the slower things go. The key to removing the contention is to not require the superblock fields in question to be locked. We do that by not marking the superblock dirty in the transaction. IOWs, we modify the incore superblock but do not modify the cached superblock buffer. In short, we do not log superblock modifications to critical fields in the superblock on every transaction. In fact we only do it just before we write the superblock to disk every sync period or just before unmount. This creates an interesting problem - if we don't log or write out the fields in every transaction, then how do the values get recovered after a crash? the answer is simple - we keep enough duplicate, logged information in other structures that we can reconstruct the correct count after log recovery has been performed. It is the AGF and AGI structures that contain the duplicate information; after recovery, we walk every AGI and AGF and sum their individual counters to get the correct value, and we do a transaction into the log to correct them. An optimisation of this is that if we have a clean unmount record, we know the value in the superblock is correct, so we can avoid the summation walk under normal conditions and so mount/recovery times do not change under normal operation. One wrinkle that was discovered during development was that the blocks used in the freespace btrees are never accounted for in the AGF counters. This was once a valid optimisation to make; when the filesystem is full, the free space btrees are empty and consume no space. Hence when it matters, the "accounting" is correct. But that means the when we do the AGF summations, we would not have a correct count and xfs_check would complain. Hence a new counter was added to track the number of blocks used by the free space btrees. This is an *on-disk format change*. As a result of this, lazy superblock counters are a mkfs option and at the moment on linux there is no way to convert an old filesystem. This is possible - xfs_db can be used to twiddle the right bits and then xfs_repair will do the format conversion for you. Similarly, you can convert backwards as well. At some point we'll add functionality to xfs_admin to do the bit twiddling easily.... SGI-PV: 964999 SGI-Modid: xfs-linux-melb:xfs-kern:28652a Signed-off-by: David Chinner <dgc@sgi.com> Signed-off-by: Christoph Hellwig <hch@infradead.org> Signed-off-by: Tim Shimmin <tes@sgi.com>
2007-05-24 13:26:31 +08:00
/*
* If asked, update the disk superblock with incore counter values if we
* are using non-persistent counters so that they don't get too far out
* of sync if we crash or get a forced shutdown. We don't want to force
* this to disk, just get a transaction into the iclogs....
*/
if (flags & SYNC_SUPER)
xfs_log_sbcount(mp, 0);
/*
* Now check to see if the log needs a "dummy" transaction.
*/
if (!(flags & SYNC_REMOUNT) && xfs_log_need_covered(mp)) {
xfs_trans_t *tp;
xfs_inode_t *ip;
/*
* Put a dummy transaction in the log to tell
* recovery that all others are OK.
*/
tp = xfs_trans_alloc(mp, XFS_TRANS_DUMMY1);
if ((error = xfs_trans_reserve(tp, 0,
XFS_ICHANGE_LOG_RES(mp),
0, 0, 0))) {
xfs_trans_cancel(tp, 0);
return error;
}
ip = mp->m_rootip;
xfs_ilock(ip, XFS_ILOCK_EXCL);
xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_ihold(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
error = xfs_trans_commit(tp, 0);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
xfs_log_force(mp, (xfs_lsn_t)0, log_flags);
}
/*
* When shutting down, we need to insure that the AIL is pushed
* to disk or the filesystem can appear corrupt from the PROM.
*/
if ((flags & (SYNC_CLOSE|SYNC_WAIT)) == (SYNC_CLOSE|SYNC_WAIT)) {
XFS_bflush(mp->m_ddev_targp);
if (mp->m_rtdev_targp) {
XFS_bflush(mp->m_rtdev_targp);
}
}
return XFS_ERROR(last_error);
}