#define XFS_BLF_AGF_BUF (1<<6)
#define XFS_BLF_AGFL_BUF (1<<7)
#define XFS_BLF_AGI_BUF (1<<8)
+#define XFS_BLF_DINO_BUF (1<<9)
#define XFS_BLF_TYPE_MASK \
(XFS_BLF_UDQUOT_BUF | \
XFS_BLF_BTREE_BUF | \
XFS_BLF_AGF_BUF | \
XFS_BLF_AGFL_BUF | \
- XFS_BLF_AGI_BUF)
+ XFS_BLF_AGI_BUF | \
+ XFS_BLF_DINO_BUF)
#define XFS_BLF_CHUNK 128
#define XFS_BLF_SHIFT 7
#define __XFS_DINODE_H__
#define XFS_DINODE_MAGIC 0x494e /* 'IN' */
-#define XFS_DINODE_GOOD_VERSION(v) (((v) == 1 || (v) == 2))
+#define XFS_DINODE_GOOD_VERSION(v) ((v) >= 1 && (v) <= 3)
typedef struct xfs_timestamp {
__be32 t_sec; /* timestamp seconds */
/* di_next_unlinked is the only non-core field in the old dinode */
__be32 di_next_unlinked;/* agi unlinked list ptr */
-} __attribute__((packed)) xfs_dinode_t;
+
+ /* start of the extended dinode, writable fields */
+ __le32 di_crc; /* CRC of the inode */
+ __be64 di_changecount; /* number of attribute changes */
+ __be64 di_lsn; /* flush sequence */
+ __be64 di_flags2; /* more random flags */
+ __u8 di_pad2[16]; /* more padding for future expansion */
+
+ /* fields only written to during inode creation */
+ xfs_timestamp_t di_crtime; /* time created */
+ __be64 di_ino; /* inode number */
+ uuid_t di_uuid; /* UUID of the filesystem */
+
+ /* structure must be padded to 64 bit alignment */
+} xfs_dinode_t;
#define DI_MAX_FLUSH 0xffff
/*
+ * Size of the core inode on disk. Version 1 and 2 inodes have
+ * the same size, but version 3 has grown a few additional fields.
+ */
+static inline uint xfs_dinode_size(int version)
+{
+ if (version == 3)
+ return sizeof(struct xfs_dinode);
+ return offsetof(struct xfs_dinode, di_crc);
+}
+
+/*
* The 32 bit link count in the inode theoretically maxes out at UINT_MAX.
* Since the pathconf interface is signed, we use 2^31 - 1 instead.
* The old inode format had a 16 bit link count, so its maximum is USHRT_MAX.
* Inode size for given fs.
*/
#define XFS_LITINO(mp, version) \
- ((int)(((mp)->m_sb.sb_inodesize) - sizeof(struct xfs_dinode)))
+ ((int)(((mp)->m_sb.sb_inodesize) - xfs_dinode_size(version)))
#define XFS_BROOT_SIZE_ADJ(ip) \
(XFS_BMBT_BLOCK_LEN((ip)->i_mount) - sizeof(xfs_bmdr_block_t))
* Return pointers to the data or attribute forks.
*/
#define XFS_DFORK_DPTR(dip) \
- ((char *)(dip) + sizeof(struct xfs_dinode))
+ ((char *)dip + xfs_dinode_size(dip->di_version))
#define XFS_DFORK_APTR(dip) \
(XFS_DFORK_DPTR(dip) + XFS_DFORK_BOFF(dip))
#define XFS_DFORK_PTR(dip,w) \
int version;
int i, j;
xfs_daddr_t d;
+ xfs_ino_t ino = 0;
/*
* Loop over the new block(s), filling in the inodes.
}
/*
- * Figure out what version number to use in the inodes we create.
- * If the superblock version has caught up to the one that supports
- * the new inode format, then use the new inode version. Otherwise
- * use the old version so that old kernels will continue to be
- * able to use the file system.
+ * Figure out what version number to use in the inodes we create. If
+ * the superblock version has caught up to the one that supports the new
+ * inode format, then use the new inode version. Otherwise use the old
+ * version so that old kernels will continue to be able to use the file
+ * system.
+ *
+ * For v3 inodes, we also need to write the inode number into the inode,
+ * so calculate the first inode number of the chunk here as
+ * XFS_OFFBNO_TO_AGINO() only works within a filesystem block, not
+ * across multiple filesystem blocks (such as a cluster) and so cannot
+ * be used in the cluster buffer loop below.
+ *
+ * Further, because we are writing the inode directly into the buffer
+ * and calculating a CRC on the entire inode, we have ot log the entire
+ * inode so that the entire range the CRC covers is present in the log.
+ * That means for v3 inode we log the entire buffer rather than just the
+ * inode cores.
*/
- if (xfs_sb_version_hasnlink(&mp->m_sb))
+ if (xfs_sb_version_hascrc(&mp->m_sb)) {
+ version = 3;
+ ino = XFS_AGINO_TO_INO(mp, agno,
+ XFS_OFFBNO_TO_AGINO(mp, agbno, 0));
+ } else if (xfs_sb_version_hasnlink(&mp->m_sb))
version = 2;
else
version = 1;
* individual transactions causing a lot of log traffic.
*/
fbuf->b_ops = &xfs_inode_buf_ops;
- xfs_buf_zero(fbuf, 0, ninodes << mp->m_sb.sb_inodelog);
+ xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
for (i = 0; i < ninodes; i++) {
int ioffset = i << mp->m_sb.sb_inodelog;
- uint isize = sizeof(struct xfs_dinode);
+ uint isize = xfs_dinode_size(version);
free = xfs_make_iptr(mp, fbuf, i);
free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
free->di_version = version;
free->di_gen = cpu_to_be32(gen);
free->di_next_unlinked = cpu_to_be32(NULLAGINO);
- xfs_trans_log_buf(tp, fbuf, ioffset, ioffset + isize - 1);
+
+ if (version == 3) {
+ free->di_ino = cpu_to_be64(ino);
+ ino++;
+ uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid);
+ xfs_dinode_calc_crc(mp, free);
+ } else {
+ /* just log the inode core */
+ xfs_trans_log_buf(tp, fbuf, ioffset,
+ ioffset + isize - 1);
+ }
+ }
+ if (version == 3) {
+ /* need to log the entire buffer */
+ xfs_trans_log_buf(tp, fbuf, 0,
+ BBTOB(fbuf->b_length) - 1);
}
xfs_trans_inode_alloc_buf(tp, fbuf);
}
#include "xfs_quota.h"
#include "xfs_filestream.h"
#include "xfs_vnodeops.h"
+#include "xfs_cksum.h"
#include "xfs_trace.h"
#include "xfs_icache.h"
to->di_dmstate = be16_to_cpu(from->di_dmstate);
to->di_flags = be16_to_cpu(from->di_flags);
to->di_gen = be32_to_cpu(from->di_gen);
+
+ if (to->di_version == 3) {
+ to->di_changecount = be64_to_cpu(from->di_changecount);
+ to->di_crtime.t_sec = be32_to_cpu(from->di_crtime.t_sec);
+ to->di_crtime.t_nsec = be32_to_cpu(from->di_crtime.t_nsec);
+ to->di_flags2 = be64_to_cpu(from->di_flags2);
+ to->di_ino = be64_to_cpu(from->di_ino);
+ to->di_lsn = be64_to_cpu(from->di_lsn);
+ memcpy(to->di_pad2, from->di_pad2, sizeof(to->di_pad2));
+ uuid_copy(&to->di_uuid, &from->di_uuid);
+ }
}
void
to->di_dmstate = cpu_to_be16(from->di_dmstate);
to->di_flags = cpu_to_be16(from->di_flags);
to->di_gen = cpu_to_be32(from->di_gen);
+
+ if (from->di_version == 3) {
+ to->di_changecount = cpu_to_be64(from->di_changecount);
+ to->di_crtime.t_sec = cpu_to_be32(from->di_crtime.t_sec);
+ to->di_crtime.t_nsec = cpu_to_be32(from->di_crtime.t_nsec);
+ to->di_flags2 = cpu_to_be64(from->di_flags2);
+ to->di_ino = cpu_to_be64(from->di_ino);
+ to->di_lsn = cpu_to_be64(from->di_lsn);
+ memcpy(to->di_pad2, from->di_pad2, sizeof(to->di_pad2));
+ uuid_copy(&to->di_uuid, &from->di_uuid);
+ }
}
STATIC uint
(XFS_DFORK_Q(dip) ? XFS_XFLAG_HASATTR : 0);
}
+static bool
+xfs_dinode_verify(
+ struct xfs_mount *mp,
+ struct xfs_inode *ip,
+ struct xfs_dinode *dip)
+{
+ if (dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC))
+ return false;
+
+ /* only version 3 or greater inodes are extensively verified here */
+ if (dip->di_version < 3)
+ return true;
+
+ if (!xfs_sb_version_hascrc(&mp->m_sb))
+ return false;
+ if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize,
+ offsetof(struct xfs_dinode, di_crc)))
+ return false;
+ if (be64_to_cpu(dip->di_ino) != ip->i_ino)
+ return false;
+ if (!uuid_equal(&dip->di_uuid, &mp->m_sb.sb_uuid))
+ return false;
+ return true;
+}
+
+void
+xfs_dinode_calc_crc(
+ struct xfs_mount *mp,
+ struct xfs_dinode *dip)
+{
+ __uint32_t crc;
+
+ if (dip->di_version < 3)
+ return;
+
+ ASSERT(xfs_sb_version_hascrc(&mp->m_sb));
+ crc = xfs_start_cksum((char *)dip, mp->m_sb.sb_inodesize,
+ offsetof(struct xfs_dinode, di_crc));
+ dip->di_crc = xfs_end_cksum(crc);
+}
+
/*
* Read the disk inode attributes into the in-core inode structure.
*/
if (error)
return error;
- /*
- * If we got something that isn't an inode it means someone
- * (nfs or dmi) has a stale handle.
- */
- if (dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC)) {
-#ifdef DEBUG
- xfs_alert(mp,
- "%s: dip->di_magic (0x%x) != XFS_DINODE_MAGIC (0x%x)",
- __func__, be16_to_cpu(dip->di_magic), XFS_DINODE_MAGIC);
-#endif /* DEBUG */
- error = XFS_ERROR(EINVAL);
+ /* even unallocated inodes are verified */
+ if (!xfs_dinode_verify(mp, ip, dip)) {
+ xfs_alert(mp, "%s: validation failed for inode %lld failed",
+ __func__, ip->i_ino);
+
+ XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp, dip);
+ error = XFS_ERROR(EFSCORRUPTED);
goto out_brelse;
}
goto out_brelse;
}
} else {
+ /*
+ * Partial initialisation of the in-core inode. Just the bits
+ * that xfs_ialloc won't overwrite or relies on being correct.
+ */
ip->i_d.di_magic = be16_to_cpu(dip->di_magic);
ip->i_d.di_version = dip->di_version;
ip->i_d.di_gen = be32_to_cpu(dip->di_gen);
ip->i_d.di_flushiter = be16_to_cpu(dip->di_flushiter);
+
+ if (dip->di_version == 3) {
+ ip->i_d.di_ino = be64_to_cpu(dip->di_ino);
+ uuid_copy(&ip->i_d.di_uuid, &dip->di_uuid);
+ }
+
/*
* Make sure to pull in the mode here as well in
* case the inode is released without being used.
xfs_buf_t **ialloc_context,
xfs_inode_t **ipp)
{
+ struct xfs_mount *mp = tp->t_mountp;
xfs_ino_t ino;
xfs_inode_t *ip;
uint flags;
* This is because we're setting fields here we need
* to prevent others from looking at until we're done.
*/
- error = xfs_iget(tp->t_mountp, tp, ino, XFS_IGET_CREATE,
+ error = xfs_iget(mp, tp, ino, XFS_IGET_CREATE,
XFS_ILOCK_EXCL, &ip);
if (error)
return error;
* 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(&tp->t_mountp->m_sb) &&
+ if (xfs_sb_version_hasnlink(&mp->m_sb) &&
ip->i_d.di_version == 1) {
ip->i_d.di_version = 2;
/*
ip->i_d.di_dmevmask = 0;
ip->i_d.di_dmstate = 0;
ip->i_d.di_flags = 0;
+
+ if (ip->i_d.di_version == 3) {
+ ASSERT(ip->i_d.di_ino == ino);
+ ASSERT(uuid_equal(&ip->i_d.di_uuid, &mp->m_sb.sb_uuid));
+ ip->i_d.di_crc = 0;
+ ip->i_d.di_changecount = 1;
+ ip->i_d.di_lsn = 0;
+ ip->i_d.di_flags2 = 0;
+ memset(&(ip->i_d.di_pad2[0]), 0, sizeof(ip->i_d.di_pad2));
+ ip->i_d.di_crtime = ip->i_d.di_mtime;
+ }
+
+
flags = XFS_ILOG_CORE;
switch (mode & S_IFMT) {
case S_IFIFO:
STATIC int
xfs_iflush_int(
- xfs_inode_t *ip,
- xfs_buf_t *bp)
+ struct xfs_inode *ip,
+ struct xfs_buf *bp)
{
- xfs_inode_log_item_t *iip;
- xfs_dinode_t *dip;
- xfs_mount_t *mp;
+ struct xfs_inode_log_item *iip = ip->i_itemp;
+ struct xfs_dinode *dip;
+ struct xfs_mount *mp = ip->i_mount;
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
ASSERT(xfs_isiflocked(ip));
ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
-
- iip = ip->i_itemp;
- mp = ip->i_mount;
+ ASSERT(iip != NULL && iip->ili_fields != 0);
/* set *dip = inode's place in the buffer */
dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
}
/*
* bump the flush iteration count, used to detect flushes which
- * postdate a log record during recovery.
+ * postdate a log record during recovery. This is redundant as we now
+ * log every change and hence this can't happen. Still, it doesn't hurt.
*/
-
ip->i_d.di_flushiter++;
/*
* need the AIL lock, because it is a 64 bit value that cannot be read
* atomically.
*/
- if (iip != NULL && iip->ili_fields != 0) {
- iip->ili_last_fields = iip->ili_fields;
- iip->ili_fields = 0;
- iip->ili_logged = 1;
+ iip->ili_last_fields = iip->ili_fields;
+ iip->ili_fields = 0;
+ iip->ili_logged = 1;
- xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
- &iip->ili_item.li_lsn);
+ xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
+ &iip->ili_item.li_lsn);
- /*
- * Attach the function xfs_iflush_done to the inode's
- * buffer. This will remove the inode from the AIL
- * and unlock the inode's flush lock when the inode is
- * completely written to disk.
- */
- xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item);
+ /*
+ * Attach the function xfs_iflush_done to the inode's
+ * buffer. This will remove the inode from the AIL
+ * and unlock the inode's flush lock when the inode is
+ * completely written to disk.
+ */
+ xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item);
- ASSERT(bp->b_fspriv != NULL);
- ASSERT(bp->b_iodone != NULL);
- } else {
- /*
- * We're flushing an inode which is not in the AIL and has
- * not been logged. For this case we can immediately drop
- * the inode flush lock because we can avoid the whole
- * AIL state thing. It's OK to drop the flush lock now,
- * because we've already locked the buffer and to do anything
- * you really need both.
- */
- if (iip != NULL) {
- ASSERT(iip->ili_logged == 0);
- ASSERT(iip->ili_last_fields == 0);
- ASSERT((iip->ili_item.li_flags & XFS_LI_IN_AIL) == 0);
- }
- xfs_ifunlock(ip);
- }
+ /* update the lsn in the on disk inode if required */
+ if (ip->i_d.di_version == 3)
+ dip->di_lsn = cpu_to_be64(iip->ili_item.li_lsn);
+
+ /* generate the checksum. */
+ xfs_dinode_calc_crc(mp, dip);
+ ASSERT(bp->b_fspriv != NULL);
+ ASSERT(bp->b_iodone != NULL);
return 0;
corrupt_out:
__uint16_t di_dmstate; /* DMIG state info */
__uint16_t di_flags; /* random flags, XFS_DIFLAG_... */
__uint32_t di_gen; /* generation number */
+
+ /* di_next_unlinked is the only non-core field in the old dinode */
+ xfs_agino_t di_next_unlinked;/* agi unlinked list ptr */
+
+ /* start of the extended dinode, writable fields */
+ __uint32_t di_crc; /* CRC of the inode */
+ __uint64_t di_changecount; /* number of attribute changes */
+ xfs_lsn_t di_lsn; /* flush sequence */
+ __uint64_t di_flags2; /* more random flags */
+ __uint8_t di_pad2[16]; /* more padding for future expansion */
+
+ /* fields only written to during inode creation */
+ xfs_ictimestamp_t di_crtime; /* time created */
+ xfs_ino_t di_ino; /* inode number */
+ uuid_t di_uuid; /* UUID of the filesystem */
+
+ /* structure must be padded to 64 bit alignment */
} xfs_icdinode_t;
+static inline uint xfs_icdinode_size(int version)
+{
+ if (version == 3)
+ return sizeof(struct xfs_icdinode);
+ return offsetof(struct xfs_icdinode, di_next_unlinked);
+}
+
/*
* Flags for xfs_ichgtime().
*/
#define XFS_ICHGTIME_MOD 0x1 /* data fork modification timestamp */
#define XFS_ICHGTIME_CHG 0x2 /* inode field change timestamp */
+#define XFS_ICHGTIME_CREATE 0x4 /* inode create timestamp */
/*
* Per-fork incore inode flags.
struct xfs_buf **, uint, uint);
int xfs_iread(struct xfs_mount *, struct xfs_trans *,
struct xfs_inode *, uint);
+void xfs_dinode_calc_crc(struct xfs_mount *, struct xfs_dinode *);
void xfs_dinode_to_disk(struct xfs_dinode *,
struct xfs_icdinode *);
void xfs_idestroy_fork(struct xfs_inode *, int);
nvecs = 1;
vecp->i_addr = &ip->i_d;
- vecp->i_len = sizeof(struct xfs_icdinode);
+ vecp->i_len = xfs_icdinode_size(ip->i_d.di_version);
vecp->i_type = XLOG_REG_TYPE_ICORE;
vecp++;
nvecs++;
xfs_agino_t *buffer_nextp;
trace_xfs_log_recover_buf_inode_buf(mp->m_log, buf_f);
+ bp->b_ops = &xfs_inode_buf_ops;
inodes_per_buf = BBTOB(bp->b_io_length) >> mp->m_sb.sb_inodelog;
for (i = 0; i < inodes_per_buf; i++) {
}
bp->b_ops = &xfs_dquot_buf_ops;
break;
+ case XFS_BLF_DINO_BUF:
+ /*
+ * we get here with inode allocation buffers, not buffers that
+ * track unlinked list changes.
+ */
+ if (*(__be16 *)bp->b_addr != cpu_to_be16(XFS_DINODE_MAGIC)) {
+ xfs_warn(mp, "Bad INODE block magic!");
+ ASSERT(0);
+ break;
+ }
+ bp->b_ops = &xfs_inode_buf_ops;
+ break;
default:
break;
}
int attr_index;
uint fields;
xfs_icdinode_t *dicp;
+ uint isize;
int need_free = 0;
if (item->ri_buf[0].i_len == sizeof(xfs_inode_log_format_t)) {
trace_xfs_log_recover_inode_recover(log, in_f);
bp = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len, 0,
- NULL);
+ &xfs_inode_buf_ops);
if (!bp) {
error = ENOMEM;
goto error;
error = EFSCORRUPTED;
goto error;
}
- if (unlikely(item->ri_buf[1].i_len > sizeof(struct xfs_icdinode))) {
+ isize = xfs_icdinode_size(dicp->di_version);
+ if (unlikely(item->ri_buf[1].i_len > isize)) {
XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(7)",
XFS_ERRLEVEL_LOW, mp, dicp);
xfs_buf_relse(bp);
}
/* The core is in in-core format */
- xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr);
+ xfs_dinode_to_disk(dip, dicp);
/* the rest is in on-disk format */
- if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
- memcpy((xfs_caddr_t) dip + sizeof(struct xfs_icdinode),
- item->ri_buf[1].i_addr + sizeof(struct xfs_icdinode),
- item->ri_buf[1].i_len - sizeof(struct xfs_icdinode));
+ if (item->ri_buf[1].i_len > isize) {
+ memcpy((char *)dip + isize,
+ item->ri_buf[1].i_addr + isize,
+ item->ri_buf[1].i_len - isize);
}
fields = in_f->ilf_fields;
}
write_inode_buffer:
+ /* re-generate the checksum. */
+ xfs_dinode_calc_crc(log->l_mp, dip);
+
ASSERT(bp->b_target->bt_mount == mp);
bp->b_iodone = xlog_recover_iodone;
xfs_buf_delwri_queue(bp, buffer_list);
ASSERT(atomic_read(&bip->bli_refcount) > 0);
bip->bli_flags |= XFS_BLI_INODE_BUF;
+ xfs_trans_buf_set_type(tp, bp, XFS_BLF_DINO_BUF);
}
/*
* This call is used to indicate that the buffer is going to
* be staled and was an inode buffer. This means it gets
- * special processing during unpin - where any inodes
+ * special processing during unpin - where any inodes
* associated with the buffer should be removed from ail.
* There is also special processing during recovery,
* any replay of the inodes in the buffer needs to be
bip->bli_flags |= XFS_BLI_STALE_INODE;
bip->bli_item.li_cb = xfs_buf_iodone;
+ xfs_trans_buf_set_type(tp, bp, XFS_BLF_DINO_BUF);
}
/*
ASSERT(atomic_read(&bip->bli_refcount) > 0);
bip->bli_flags |= XFS_BLI_INODE_ALLOC_BUF;
+ xfs_trans_buf_set_type(tp, bp, XFS_BLF_DINO_BUF);
}
/*
projects / profile / common / kernel-common.git / commitdiff