aboutsummaryrefslogtreecommitdiff
path: root/fs/xfs/xfs_inode.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/xfs/xfs_inode.c')
-rw-r--r--fs/xfs/xfs_inode.c719
1 files changed, 386 insertions, 333 deletions
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index e3d75385aa7..a6115fe1ac9 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -19,29 +19,24 @@
#include "xfs.h"
#include "xfs_fs.h"
+#include "xfs_shared.h"
#include "xfs_format.h"
-#include "xfs_log.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_trans_space.h"
-#include "xfs_trans_priv.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_da_format.h"
#include "xfs_da_btree.h"
-#include "xfs_dir2_format.h"
#include "xfs_dir2.h"
-#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
#include "xfs_attr_sf.h"
#include "xfs_attr.h"
-#include "xfs_dinode.h"
-#include "xfs_inode.h"
+#include "xfs_trans_space.h"
+#include "xfs_trans.h"
#include "xfs_buf_item.h"
#include "xfs_inode_item.h"
-#include "xfs_btree.h"
-#include "xfs_alloc.h"
#include "xfs_ialloc.h"
#include "xfs_bmap.h"
#include "xfs_bmap_util.h"
@@ -52,6 +47,9 @@
#include "xfs_trace.h"
#include "xfs_icache.h"
#include "xfs_symlink.h"
+#include "xfs_trans_priv.h"
+#include "xfs_log.h"
+#include "xfs_bmap_btree.h"
kmem_zone_t *xfs_inode_zone;
@@ -63,6 +61,8 @@ kmem_zone_t *xfs_inode_zone;
STATIC int xfs_iflush_int(xfs_inode_t *, xfs_buf_t *);
+STATIC int xfs_iunlink_remove(xfs_trans_t *, xfs_inode_t *);
+
/*
* helper function to extract extent size hint from inode
*/
@@ -78,48 +78,44 @@ xfs_get_extsz_hint(
}
/*
- * This is a wrapper routine around the xfs_ilock() routine used to centralize
- * some grungy code. It is used in places that wish to lock the inode solely
- * for reading the extents. The reason these places can't just call
- * xfs_ilock(SHARED) is that the inode lock also guards to bringing in of the
- * extents from disk for a file in b-tree format. If the inode is in b-tree
- * format, then we need to lock the inode exclusively until the extents are read
- * in. Locking it exclusively all the time would limit our parallelism
- * unnecessarily, though. What we do instead is check to see if the extents
- * have been read in yet, and only lock the inode exclusively if they have not.
+ * These two are wrapper routines around the xfs_ilock() routine used to
+ * centralize some grungy code. They are used in places that wish to lock the
+ * inode solely for reading the extents. The reason these places can't just
+ * call xfs_ilock(ip, XFS_ILOCK_SHARED) is that the inode lock also guards to
+ * bringing in of the extents from disk for a file in b-tree format. If the
+ * inode is in b-tree format, then we need to lock the inode exclusively until
+ * the extents are read in. Locking it exclusively all the time would limit
+ * our parallelism unnecessarily, though. What we do instead is check to see
+ * if the extents have been read in yet, and only lock the inode exclusively
+ * if they have not.
*
- * The function returns a value which should be given to the corresponding
- * xfs_iunlock_map_shared(). This value is the mode in which the lock was
- * actually taken.
+ * The functions return a value which should be given to the corresponding
+ * xfs_iunlock() call.
*/
uint
-xfs_ilock_map_shared(
- xfs_inode_t *ip)
+xfs_ilock_data_map_shared(
+ struct xfs_inode *ip)
{
- uint lock_mode;
+ uint lock_mode = XFS_ILOCK_SHARED;
- if ((ip->i_d.di_format == XFS_DINODE_FMT_BTREE) &&
- ((ip->i_df.if_flags & XFS_IFEXTENTS) == 0)) {
+ if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE &&
+ (ip->i_df.if_flags & XFS_IFEXTENTS) == 0)
lock_mode = XFS_ILOCK_EXCL;
- } else {
- lock_mode = XFS_ILOCK_SHARED;
- }
-
xfs_ilock(ip, lock_mode);
-
return lock_mode;
}
-/*
- * This is simply the unlock routine to go with xfs_ilock_map_shared().
- * All it does is call xfs_iunlock() with the given lock_mode.
- */
-void
-xfs_iunlock_map_shared(
- xfs_inode_t *ip,
- unsigned int lock_mode)
+uint
+xfs_ilock_attr_map_shared(
+ struct xfs_inode *ip)
{
- xfs_iunlock(ip, lock_mode);
+ uint lock_mode = XFS_ILOCK_SHARED;
+
+ if (ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE &&
+ (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
+ lock_mode = XFS_ILOCK_EXCL;
+ xfs_ilock(ip, lock_mode);
+ return lock_mode;
}
/*
@@ -589,9 +585,9 @@ xfs_lookup(
if (XFS_FORCED_SHUTDOWN(dp->i_mount))
return XFS_ERROR(EIO);
- lock_mode = xfs_ilock_map_shared(dp);
+ lock_mode = xfs_ilock_data_map_shared(dp);
error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
- xfs_iunlock_map_shared(dp, lock_mode);
+ xfs_iunlock(dp, lock_mode);
if (error)
goto out;
@@ -659,7 +655,6 @@ xfs_ialloc(
uint flags;
int error;
timespec_t tv;
- int filestreams = 0;
/*
* Call the space management code to pick
@@ -686,6 +681,14 @@ xfs_ialloc(
return error;
ASSERT(ip != NULL);
+ /*
+ * We always convert v1 inodes to v2 now - we only support filesystems
+ * with >= v2 inode capability, so there is no reason for ever leaving
+ * an inode in v1 format.
+ */
+ if (ip->i_d.di_version == 1)
+ ip->i_d.di_version = 2;
+
ip->i_d.di_mode = mode;
ip->i_d.di_onlink = 0;
ip->i_d.di_nlink = nlink;
@@ -695,27 +698,6 @@ xfs_ialloc(
xfs_set_projid(ip, prid);
memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
- /*
- * If the superblock version is up to where we support new format
- * inodes and this is currently an old format inode, then change
- * the inode version number now. This way we only do the conversion
- * here rather than here and in the flush/logging code.
- */
- if (xfs_sb_version_hasnlink(&mp->m_sb) &&
- ip->i_d.di_version == 1) {
- ip->i_d.di_version = 2;
- /*
- * We've already zeroed the old link count, the projid field,
- * and the pad field.
- */
- }
-
- /*
- * Project ids won't be stored on disk if we are using a version 1 inode.
- */
- if ((prid != 0) && (ip->i_d.di_version == 1))
- xfs_bump_ino_vers2(tp, ip);
-
if (pip && XFS_INHERIT_GID(pip)) {
ip->i_d.di_gid = pip->i_d.di_gid;
if ((pip->i_d.di_mode & S_ISGID) && S_ISDIR(mode)) {
@@ -776,13 +758,6 @@ xfs_ialloc(
flags |= XFS_ILOG_DEV;
break;
case S_IFREG:
- /*
- * we can't set up filestreams until after the VFS inode
- * is set up properly.
- */
- if (pip && xfs_inode_is_filestream(pip))
- filestreams = 1;
- /* fall through */
case S_IFDIR:
if (pip && (pip->i_d.di_flags & XFS_DIFLAG_ANY)) {
uint di_flags = 0;
@@ -848,15 +823,6 @@ xfs_ialloc(
/* now that we have an i_mode we can setup inode ops and unlock */
xfs_setup_inode(ip);
- /* now we have set up the vfs inode we can associate the filestream */
- if (filestreams) {
- error = xfs_filestream_associate(pip, ip);
- if (error < 0)
- return -error;
- if (!error)
- xfs_iflags_set(ip, XFS_IFILESTREAM);
- }
-
*ipp = ip;
return 0;
}
@@ -1077,40 +1043,6 @@ xfs_droplink(
}
/*
- * This gets called when the inode's version needs to be changed from 1 to 2.
- * Currently this happens when the nlink field overflows the old 16-bit value
- * or when chproj is called to change the project for the first time.
- * As a side effect the superblock version will also get rev'd
- * to contain the NLINK bit.
- */
-void
-xfs_bump_ino_vers2(
- xfs_trans_t *tp,
- xfs_inode_t *ip)
-{
- xfs_mount_t *mp;
-
- ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
- ASSERT(ip->i_d.di_version == 1);
-
- ip->i_d.di_version = 2;
- ip->i_d.di_onlink = 0;
- memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
- mp = tp->t_mountp;
- if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
- spin_lock(&mp->m_sb_lock);
- if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
- xfs_sb_version_addnlink(&mp->m_sb);
- spin_unlock(&mp->m_sb_lock);
- xfs_mod_sb(tp, XFS_SB_VERSIONNUM);
- } else {
- spin_unlock(&mp->m_sb_lock);
- }
- }
- /* Caller must log the inode */
-}
-
-/*
* Increment the link count on an inode & log the change.
*/
int
@@ -1120,22 +1052,10 @@ xfs_bumplink(
{
xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
- ASSERT(ip->i_d.di_nlink > 0);
+ ASSERT(ip->i_d.di_version > 1);
+ ASSERT(ip->i_d.di_nlink > 0 || (VFS_I(ip)->i_state & I_LINKABLE));
ip->i_d.di_nlink++;
inc_nlink(VFS_I(ip));
- if ((ip->i_d.di_version == 1) &&
- (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
- /*
- * The inode has increased its number of links beyond
- * what can fit in an old format inode. It now needs
- * to be converted to a version 2 inode with a 32 bit
- * link count. If this is the first inode in the file
- * system to do this, then we need to bump the superblock
- * version number as well.
- */
- xfs_bump_ino_vers2(tp, ip);
- }
-
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
return 0;
}
@@ -1170,10 +1090,7 @@ xfs_create(
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
- prid = xfs_get_projid(dp);
- else
- prid = XFS_PROJID_DEFAULT;
+ prid = xfs_get_initial_prid(dp);
/*
* Make sure that we have allocated dquot(s) on disk.
@@ -1338,6 +1255,114 @@ xfs_create(
}
int
+xfs_create_tmpfile(
+ struct xfs_inode *dp,
+ struct dentry *dentry,
+ umode_t mode,
+ struct xfs_inode **ipp)
+{
+ struct xfs_mount *mp = dp->i_mount;
+ struct xfs_inode *ip = NULL;
+ struct xfs_trans *tp = NULL;
+ int error;
+ uint cancel_flags = XFS_TRANS_RELEASE_LOG_RES;
+ prid_t prid;
+ struct xfs_dquot *udqp = NULL;
+ struct xfs_dquot *gdqp = NULL;
+ struct xfs_dquot *pdqp = NULL;
+ struct xfs_trans_res *tres;
+ uint resblks;
+
+ if (XFS_FORCED_SHUTDOWN(mp))
+ return XFS_ERROR(EIO);
+
+ prid = xfs_get_initial_prid(dp);
+
+ /*
+ * Make sure that we have allocated dquot(s) on disk.
+ */
+ error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()),
+ xfs_kgid_to_gid(current_fsgid()), prid,
+ XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
+ &udqp, &gdqp, &pdqp);
+ if (error)
+ return error;
+
+ resblks = XFS_IALLOC_SPACE_RES(mp);
+ tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE_TMPFILE);
+
+ tres = &M_RES(mp)->tr_create_tmpfile;
+ error = xfs_trans_reserve(tp, tres, resblks, 0);
+ if (error == ENOSPC) {
+ /* No space at all so try a "no-allocation" reservation */
+ resblks = 0;
+ error = xfs_trans_reserve(tp, tres, 0, 0);
+ }
+ if (error) {
+ cancel_flags = 0;
+ goto out_trans_cancel;
+ }
+
+ error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
+ pdqp, resblks, 1, 0);
+ if (error)
+ goto out_trans_cancel;
+
+ error = xfs_dir_ialloc(&tp, dp, mode, 1, 0,
+ prid, resblks > 0, &ip, NULL);
+ if (error) {
+ if (error == ENOSPC)
+ goto out_trans_cancel;
+ goto out_trans_abort;
+ }
+
+ if (mp->m_flags & XFS_MOUNT_WSYNC)
+ xfs_trans_set_sync(tp);
+
+ /*
+ * Attach the dquot(s) to the inodes and modify them incore.
+ * These ids of the inode couldn't have changed since the new
+ * inode has been locked ever since it was created.
+ */
+ xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
+
+ ip->i_d.di_nlink--;
+ error = xfs_iunlink(tp, ip);
+ if (error)
+ goto out_trans_abort;
+
+ error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ if (error)
+ goto out_release_inode;
+
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
+ xfs_qm_dqrele(pdqp);
+
+ *ipp = ip;
+ return 0;
+
+ out_trans_abort:
+ cancel_flags |= XFS_TRANS_ABORT;
+ out_trans_cancel:
+ xfs_trans_cancel(tp, cancel_flags);
+ out_release_inode:
+ /*
+ * Wait until after the current transaction is aborted to
+ * release the inode. This prevents recursive transactions
+ * and deadlocks from xfs_inactive.
+ */
+ if (ip)
+ IRELE(ip);
+
+ xfs_qm_dqrele(udqp);
+ xfs_qm_dqrele(gdqp);
+ xfs_qm_dqrele(pdqp);
+
+ return error;
+}
+
+int
xfs_link(
xfs_inode_t *tdp,
xfs_inode_t *sip,
@@ -1402,6 +1427,12 @@ xfs_link(
xfs_bmap_init(&free_list, &first_block);
+ if (sip->i_d.di_nlink == 0) {
+ error = xfs_iunlink_remove(tp, sip);
+ if (error)
+ goto abort_return;
+ }
+
error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
&first_block, &free_list, resblks);
if (error)
@@ -1592,16 +1623,6 @@ xfs_release(
int truncated;
/*
- * If we are using filestreams, and we have an unlinked
- * file that we are processing the last close on, then nothing
- * will be able to reopen and write to this file. Purge this
- * inode from the filestreams cache so that it doesn't delay
- * teardown of the inode.
- */
- if ((ip->i_d.di_nlink == 0) && xfs_inode_is_filestream(ip))
- xfs_filestream_deassociate(ip);
-
- /*
* If we previously truncated this file and removed old data
* in the process, we want to initiate "early" writeout on
* the last close. This is an attempt to combat the notorious
@@ -1663,6 +1684,150 @@ xfs_release(
}
/*
+ * xfs_inactive_truncate
+ *
+ * Called to perform a truncate when an inode becomes unlinked.
+ */
+STATIC int
+xfs_inactive_truncate(
+ struct xfs_inode *ip)
+{
+ struct xfs_mount *mp = ip->i_mount;
+ struct xfs_trans *tp;
+ int error;
+
+ tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
+ error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
+ if (error) {
+ ASSERT(XFS_FORCED_SHUTDOWN(mp));
+ xfs_trans_cancel(tp, 0);
+ return error;
+ }
+
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin(tp, ip, 0);
+
+ /*
+ * Log the inode size first to prevent stale data exposure in the event
+ * of a system crash before the truncate completes. See the related
+ * comment in xfs_setattr_size() for details.
+ */
+ ip->i_d.di_size = 0;
+ xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
+ error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, 0);
+ if (error)
+ goto error_trans_cancel;
+
+ ASSERT(ip->i_d.di_nextents == 0);
+
+ error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ if (error)
+ goto error_unlock;
+
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return 0;
+
+error_trans_cancel:
+ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
+error_unlock:
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return error;
+}
+
+/*
+ * xfs_inactive_ifree()
+ *
+ * Perform the inode free when an inode is unlinked.
+ */
+STATIC int
+xfs_inactive_ifree(
+ struct xfs_inode *ip)
+{
+ xfs_bmap_free_t free_list;
+ xfs_fsblock_t first_block;
+ int committed;
+ struct xfs_mount *mp = ip->i_mount;
+ struct xfs_trans *tp;
+ int error;
+
+ tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
+
+ /*
+ * The ifree transaction might need to allocate blocks for record
+ * insertion to the finobt. We don't want to fail here at ENOSPC, so
+ * allow ifree to dip into the reserved block pool if necessary.
+ *
+ * Freeing large sets of inodes generally means freeing inode chunks,
+ * directory and file data blocks, so this should be relatively safe.
+ * Only under severe circumstances should it be possible to free enough
+ * inodes to exhaust the reserve block pool via finobt expansion while
+ * at the same time not creating free space in the filesystem.
+ *
+ * Send a warning if the reservation does happen to fail, as the inode
+ * now remains allocated and sits on the unlinked list until the fs is
+ * repaired.
+ */
+ tp->t_flags |= XFS_TRANS_RESERVE;
+ error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree,
+ XFS_IFREE_SPACE_RES(mp), 0);
+ if (error) {
+ if (error == ENOSPC) {
+ xfs_warn_ratelimited(mp,
+ "Failed to remove inode(s) from unlinked list. "
+ "Please free space, unmount and run xfs_repair.");
+ } else {
+ ASSERT(XFS_FORCED_SHUTDOWN(mp));
+ }
+ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES);
+ return error;
+ }
+
+ xfs_ilock(ip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin(tp, ip, 0);
+
+ xfs_bmap_init(&free_list, &first_block);
+ error = xfs_ifree(tp, ip, &free_list);
+ if (error) {
+ /*
+ * If we fail to free the inode, shut down. The cancel
+ * might do that, we need to make sure. Otherwise the
+ * inode might be lost for a long time or forever.
+ */
+ if (!XFS_FORCED_SHUTDOWN(mp)) {
+ xfs_notice(mp, "%s: xfs_ifree returned error %d",
+ __func__, error);
+ xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+ }
+ xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return error;
+ }
+
+ /*
+ * Credit the quota account(s). The inode is gone.
+ */
+ xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, -1);
+
+ /*
+ * Just ignore errors at this point. There is nothing we can
+ * do except to try to keep going. Make sure it's not a silent
+ * error.
+ */
+ error = xfs_bmap_finish(&tp, &free_list, &committed);
+ if (error)
+ xfs_notice(mp, "%s: xfs_bmap_finish returned error %d",
+ __func__, error);
+ error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+ if (error)
+ xfs_notice(mp, "%s: xfs_trans_commit returned error %d",
+ __func__, error);
+
+ xfs_iunlock(ip, XFS_ILOCK_EXCL);
+ return 0;
+}
+
+/*
* xfs_inactive
*
* This is called when the vnode reference count for the vnode
@@ -1670,16 +1835,11 @@ xfs_release(
* now be truncated. Also, we clear all of the read-ahead state
* kept for the inode here since the file is now closed.
*/
-int
+void
xfs_inactive(
xfs_inode_t *ip)
{
- xfs_bmap_free_t free_list;
- xfs_fsblock_t first_block;
- int committed;
- struct xfs_trans *tp;
struct xfs_mount *mp;
- struct xfs_trans_res *resp;
int error;
int truncate = 0;
@@ -1687,19 +1847,17 @@ xfs_inactive(
* If the inode is already free, then there can be nothing
* to clean up here.
*/
- if (ip->i_d.di_mode == 0 || is_bad_inode(VFS_I(ip))) {
+ if (ip->i_d.di_mode == 0) {
ASSERT(ip->i_df.if_real_bytes == 0);
ASSERT(ip->i_df.if_broot_bytes == 0);
- return VN_INACTIVE_CACHE;
+ return;
}
mp = ip->i_mount;
- error = 0;
-
/* If this is a read-only mount, don't do this (would generate I/O) */
if (mp->m_flags & XFS_MOUNT_RDONLY)
- goto out;
+ return;
if (ip->i_d.di_nlink != 0) {
/*
@@ -1707,12 +1865,10 @@ xfs_inactive(
* cache. Post-eof blocks must be freed, lest we end up with
* broken free space accounting.
*/
- if (xfs_can_free_eofblocks(ip, true)) {
- error = xfs_free_eofblocks(mp, ip, false);
- if (error)
- return VN_INACTIVE_CACHE;
- }
- goto out;
+ if (xfs_can_free_eofblocks(ip, true))
+ xfs_free_eofblocks(mp, ip, false);
+
+ return;
}
if (S_ISREG(ip->i_d.di_mode) &&
@@ -1722,36 +1878,14 @@ xfs_inactive(
error = xfs_qm_dqattach(ip, 0);
if (error)
- return VN_INACTIVE_CACHE;
-
- tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
- resp = (truncate || S_ISLNK(ip->i_d.di_mode)) ?
- &M_RES(mp)->tr_itruncate : &M_RES(mp)->tr_ifree;
-
- error = xfs_trans_reserve(tp, resp, 0, 0);
- if (error) {
- ASSERT(XFS_FORCED_SHUTDOWN(mp));
- xfs_trans_cancel(tp, 0);
- return VN_INACTIVE_CACHE;
- }
-
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, 0);
-
- if (S_ISLNK(ip->i_d.di_mode)) {
- error = xfs_inactive_symlink(ip, &tp);
- if (error)
- goto out_cancel;
- } else if (truncate) {
- ip->i_d.di_size = 0;
- xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+ return;
- error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, 0);
- if (error)
- goto out_cancel;
-
- ASSERT(ip->i_d.di_nextents == 0);
- }
+ if (S_ISLNK(ip->i_d.di_mode))
+ error = xfs_inactive_symlink(ip);
+ else if (truncate)
+ error = xfs_inactive_truncate(ip);
+ if (error)
+ return;
/*
* If there are attributes associated with the file then blow them away
@@ -1762,25 +1896,9 @@ xfs_inactive(
if (ip->i_d.di_anextents > 0) {
ASSERT(ip->i_d.di_forkoff != 0);
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error)
- goto out_unlock;
-
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
-
error = xfs_attr_inactive(ip);
if (error)
- goto out;
-
- tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
- error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree, 0, 0);
- if (error) {
- xfs_trans_cancel(tp, 0);
- goto out;
- }
-
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, 0);
+ return;
}
if (ip->i_afp)
@@ -1791,52 +1909,14 @@ xfs_inactive(
/*
* Free the inode.
*/
- xfs_bmap_init(&free_list, &first_block);
- error = xfs_ifree(tp, ip, &free_list);
- if (error) {
- /*
- * If we fail to free the inode, shut down. The cancel
- * might do that, we need to make sure. Otherwise the
- * inode might be lost for a long time or forever.
- */
- if (!XFS_FORCED_SHUTDOWN(mp)) {
- xfs_notice(mp, "%s: xfs_ifree returned error %d",
- __func__, error);
- xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
- }
- xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
- } else {
- /*
- * Credit the quota account(s). The inode is gone.
- */
- xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, -1);
-
- /*
- * Just ignore errors at this point. There is nothing we can
- * do except to try to keep going. Make sure it's not a silent
- * error.
- */
- error = xfs_bmap_finish(&tp, &free_list, &committed);
- if (error)
- xfs_notice(mp, "%s: xfs_bmap_finish returned error %d",
- __func__, error);
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error)
- xfs_notice(mp, "%s: xfs_trans_commit returned error %d",
- __func__, error);
- }
+ error = xfs_inactive_ifree(ip);
+ if (error)
+ return;
/*
* Release the dquots held by inode, if any.
*/
xfs_qm_dqdetach(ip);
-out_unlock:
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
-out:
- return VN_INACTIVE_CACHE;
-out_cancel:
- xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
- goto out_unlock;
}
/*
@@ -2107,8 +2187,8 @@ xfs_ifree_cluster(
{
xfs_mount_t *mp = free_ip->i_mount;
int blks_per_cluster;
+ int inodes_per_cluster;
int nbufs;
- int ninodes;
int i, j;
xfs_daddr_t blkno;
xfs_buf_t *bp;
@@ -2118,18 +2198,11 @@ xfs_ifree_cluster(
struct xfs_perag *pag;
pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
- if (mp->m_sb.sb_blocksize >= XFS_INODE_CLUSTER_SIZE(mp)) {
- blks_per_cluster = 1;
- ninodes = mp->m_sb.sb_inopblock;
- nbufs = XFS_IALLOC_BLOCKS(mp);
- } else {
- blks_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) /
- mp->m_sb.sb_blocksize;
- ninodes = blks_per_cluster * mp->m_sb.sb_inopblock;
- nbufs = XFS_IALLOC_BLOCKS(mp) / blks_per_cluster;
- }
+ blks_per_cluster = xfs_icluster_size_fsb(mp);
+ inodes_per_cluster = blks_per_cluster << mp->m_sb.sb_inopblog;
+ nbufs = mp->m_ialloc_blks / blks_per_cluster;
- for (j = 0; j < nbufs; j++, inum += ninodes) {
+ for (j = 0; j < nbufs; j++, inum += inodes_per_cluster) {
blkno = XFS_AGB_TO_DADDR(mp, XFS_INO_TO_AGNO(mp, inum),
XFS_INO_TO_AGBNO(mp, inum));
@@ -2191,7 +2264,7 @@ xfs_ifree_cluster(
* transaction stale above, which means there is no point in
* even trying to lock them.
*/
- for (i = 0; i < ninodes; i++) {
+ for (i = 0; i < inodes_per_cluster; i++) {
retry:
rcu_read_lock();
ip = radix_tree_lookup(&pag->pag_ici_root,
@@ -2370,6 +2443,33 @@ xfs_iunpin_wait(
__xfs_iunpin_wait(ip);
}
+/*
+ * Removing an inode from the namespace involves removing the directory entry
+ * and dropping the link count on the inode. Removing the directory entry can
+ * result in locking an AGF (directory blocks were freed) and removing a link
+ * count can result in placing the inode on an unlinked list which results in
+ * locking an AGI.
+ *
+ * The big problem here is that we have an ordering constraint on AGF and AGI
+ * locking - inode allocation locks the AGI, then can allocate a new extent for
+ * new inodes, locking the AGF after the AGI. Similarly, freeing the inode
+ * removes the inode from the unlinked list, requiring that we lock the AGI
+ * first, and then freeing the inode can result in an inode chunk being freed
+ * and hence freeing disk space requiring that we lock an AGF.
+ *
+ * Hence the ordering that is imposed by other parts of the code is AGI before
+ * AGF. This means we cannot remove the directory entry before we drop the inode
+ * reference count and put it on the unlinked list as this results in a lock
+ * order of AGF then AGI, and this can deadlock against inode allocation and
+ * freeing. Therefore we must drop the link counts before we remove the
+ * directory entry.
+ *
+ * This is still safe from a transactional point of view - it is not until we
+ * get to xfs_bmap_finish() that we have the possibility of multiple
+ * transactions in this operation. Hence as long as we remove the directory
+ * entry and drop the link count in the first transaction of the remove
+ * operation, there are no transactional constraints on the ordering here.
+ */
int
xfs_remove(
xfs_inode_t *dp,
@@ -2439,6 +2539,7 @@ xfs_remove(
/*
* If we're removing a directory perform some additional validation.
*/
+ cancel_flags |= XFS_TRANS_ABORT;
if (is_dir) {
ASSERT(ip->i_d.di_nlink >= 2);
if (ip->i_d.di_nlink != 2) {
@@ -2449,31 +2550,16 @@ xfs_remove(
error = XFS_ERROR(ENOTEMPTY);
goto out_trans_cancel;
}
- }
-
- xfs_bmap_init(&free_list, &first_block);
- error = xfs_dir_removename(tp, dp, name, ip->i_ino,
- &first_block, &free_list, resblks);
- if (error) {
- ASSERT(error != ENOENT);
- goto out_bmap_cancel;
- }
- xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
- if (is_dir) {
- /*
- * Drop the link from ip's "..".
- */
+ /* Drop the link from ip's "..". */
error = xfs_droplink(tp, dp);
if (error)
- goto out_bmap_cancel;
+ goto out_trans_cancel;
- /*
- * Drop the "." link from ip to self.
- */
+ /* Drop the "." link from ip to self. */
error = xfs_droplink(tp, ip);
if (error)
- goto out_bmap_cancel;
+ goto out_trans_cancel;
} else {
/*
* When removing a non-directory we need to log the parent
@@ -2482,20 +2568,24 @@ xfs_remove(
*/
xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
}
+ xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
- /*
- * Drop the link from dp to ip.
- */
+ /* Drop the link from dp to ip. */
error = xfs_droplink(tp, ip);
if (error)
- goto out_bmap_cancel;
+ goto out_trans_cancel;
- /*
- * Determine if this is the last link while
- * we are in the transaction.
- */
+ /* Determine if this is the last link while the inode is locked */
link_zero = (ip->i_d.di_nlink == 0);
+ xfs_bmap_init(&free_list, &first_block);
+ error = xfs_dir_removename(tp, dp, name, ip->i_ino,
+ &first_block, &free_list, resblks);
+ if (error) {
+ ASSERT(error != ENOENT);
+ goto out_bmap_cancel;
+ }
+
/*
* If this is a synchronous mount, make sure that the
* remove transaction goes to disk before returning to
@@ -2512,20 +2602,13 @@ xfs_remove(
if (error)
goto std_return;
- /*
- * If we are using filestreams, kill the stream association.
- * If the file is still open it may get a new one but that
- * will get killed on last close in xfs_close() so we don't
- * have to worry about that.
- */
- if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
+ if (is_dir && xfs_inode_is_filestream(ip))
xfs_filestream_deassociate(ip);
return 0;
out_bmap_cancel:
xfs_bmap_cancel(&free_list);
- cancel_flags |= XFS_TRANS_ABORT;
out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
std_return:
@@ -2856,13 +2939,13 @@ xfs_iflush_cluster(
pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
- inodes_per_cluster = XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog;
+ inodes_per_cluster = mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog;
ilist_size = inodes_per_cluster * sizeof(xfs_inode_t *);
ilist = kmem_alloc(ilist_size, KM_MAYFAIL|KM_NOFS);
if (!ilist)
goto out_put;
- mask = ~(((XFS_INODE_CLUSTER_SIZE(mp) >> mp->m_sb.sb_inodelog)) - 1);
+ mask = ~(((mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog)) - 1);
first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask;
rcu_read_lock();
/* really need a gang lookup range call here */
@@ -3107,6 +3190,7 @@ xfs_iflush_int(
ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
ASSERT(iip != NULL && iip->ili_fields != 0);
+ ASSERT(ip->i_d.di_version > 1);
/* set *dip = inode's place in the buffer */
dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
@@ -3167,7 +3251,7 @@ xfs_iflush_int(
}
/*
- * Inode item log recovery for v1/v2 inodes are dependent on the
+ * Inode item log recovery for v2 inodes are dependent on the
* di_flushiter count for correct sequencing. We bump the flush
* iteration count so we can detect flushes which postdate a log record
* during recovery. This is redundant as we now log every change and
@@ -3190,40 +3274,9 @@ xfs_iflush_int(
if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
ip->i_d.di_flushiter = 0;
- /*
- * If this is really an old format inode and the superblock version
- * has not been updated to support only new format inodes, then
- * convert back to the old inode format. If the superblock version
- * has been updated, then make the conversion permanent.
- */
- ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
- if (ip->i_d.di_version == 1) {
- if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
- /*
- * Convert it back.
- */
- ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
- dip->di_onlink = cpu_to_be16(ip->i_d.di_nlink);
- } else {
- /*
- * The superblock version has already been bumped,
- * so just make the conversion to the new inode
- * format permanent.
- */
- ip->i_d.di_version = 2;
- dip->di_version = 2;
- ip->i_d.di_onlink = 0;
- dip->di_onlink = 0;
- memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
- memset(&(dip->di_pad[0]), 0,
- sizeof(dip->di_pad));
- ASSERT(xfs_get_projid(ip) == 0);
- }
- }
-
- xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK, bp);
+ xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK);
if (XFS_IFORK_Q(ip))
- xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK, bp);
+ xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK);
xfs_inobp_check(mp, bp);
/*
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-10 06:40:32 +0000