diff options
Diffstat (limited to 'fs/xfs/xfs_inode.c')
| -rw-r--r-- | fs/xfs/xfs_inode.c | 5059 |
1 files changed, 2138 insertions, 2921 deletions
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c index 108c7a085f9..a6115fe1ac9 100644 --- a/fs/xfs/xfs_inode.c +++ b/fs/xfs/xfs_inode.c @@ -19,704 +19,490 @@ #include "xfs.h" #include "xfs_fs.h" -#include "xfs_types.h" -#include "xfs_bit.h" -#include "xfs_log.h" +#include "xfs_shared.h" +#include "xfs_format.h" +#include "xfs_log_format.h" +#include "xfs_trans_resv.h" #include "xfs_inum.h" -#include "xfs_trans.h" -#include "xfs_trans_priv.h" #include "xfs_sb.h" #include "xfs_ag.h" #include "xfs_mount.h" -#include "xfs_bmap_btree.h" -#include "xfs_alloc_btree.h" -#include "xfs_ialloc_btree.h" -#include "xfs_attr_sf.h" -#include "xfs_dinode.h" #include "xfs_inode.h" +#include "xfs_da_format.h" +#include "xfs_da_btree.h" +#include "xfs_dir2.h" +#include "xfs_attr_sf.h" +#include "xfs_attr.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_btree_trace.h" -#include "xfs_alloc.h" #include "xfs_ialloc.h" #include "xfs_bmap.h" +#include "xfs_bmap_util.h" #include "xfs_error.h" -#include "xfs_utils.h" #include "xfs_quota.h" #include "xfs_filestream.h" -#include "xfs_vnodeops.h" +#include "xfs_cksum.h" #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_ifork_zone; kmem_zone_t *xfs_inode_zone; /* - * Used in xfs_itruncate(). This is the maximum number of extents + * Used in xfs_itruncate_extents(). This is the maximum number of extents * freed from a file in a single transaction. */ #define XFS_ITRUNC_MAX_EXTENTS 2 STATIC int xfs_iflush_int(xfs_inode_t *, xfs_buf_t *); -STATIC int xfs_iformat_local(xfs_inode_t *, xfs_dinode_t *, int, int); -STATIC int xfs_iformat_extents(xfs_inode_t *, xfs_dinode_t *, int); -STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int); -#ifdef DEBUG +STATIC int xfs_iunlink_remove(xfs_trans_t *, xfs_inode_t *); + /* - * Make sure that the extents in the given memory buffer - * are valid. + * helper function to extract extent size hint from inode */ -STATIC void -xfs_validate_extents( - xfs_ifork_t *ifp, - int nrecs, - xfs_exntfmt_t fmt) +xfs_extlen_t +xfs_get_extsz_hint( + struct xfs_inode *ip) { - xfs_bmbt_irec_t irec; - xfs_bmbt_rec_host_t rec; - int i; - - for (i = 0; i < nrecs; i++) { - xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i); - rec.l0 = get_unaligned(&ep->l0); - rec.l1 = get_unaligned(&ep->l1); - xfs_bmbt_get_all(&rec, &irec); - if (fmt == XFS_EXTFMT_NOSTATE) - ASSERT(irec.br_state == XFS_EXT_NORM); - } + if ((ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE) && ip->i_d.di_extsize) + return ip->i_d.di_extsize; + if (XFS_IS_REALTIME_INODE(ip)) + return ip->i_mount->m_sb.sb_rextsize; + return 0; } -#else /* DEBUG */ -#define xfs_validate_extents(ifp, nrecs, fmt) -#endif /* DEBUG */ /* - * Check that none of the inode's in the buffer have a next - * unlinked field of 0. + * 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 functions return a value which should be given to the corresponding + * xfs_iunlock() call. */ -#if defined(DEBUG) -void -xfs_inobp_check( - xfs_mount_t *mp, - xfs_buf_t *bp) +uint +xfs_ilock_data_map_shared( + struct xfs_inode *ip) { - int i; - int j; - xfs_dinode_t *dip; + uint lock_mode = XFS_ILOCK_SHARED; - j = mp->m_inode_cluster_size >> mp->m_sb.sb_inodelog; + if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE && + (ip->i_df.if_flags & XFS_IFEXTENTS) == 0) + lock_mode = XFS_ILOCK_EXCL; + xfs_ilock(ip, lock_mode); + return lock_mode; +} - for (i = 0; i < j; i++) { - dip = (xfs_dinode_t *)xfs_buf_offset(bp, - i * mp->m_sb.sb_inodesize); - if (!dip->di_next_unlinked) { - xfs_fs_cmn_err(CE_ALERT, mp, - "Detected a bogus zero next_unlinked field in incore inode buffer 0x%p. About to pop an ASSERT.", - bp); - ASSERT(dip->di_next_unlinked); - } - } +uint +xfs_ilock_attr_map_shared( + struct xfs_inode *ip) +{ + 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; } -#endif /* - * Find the buffer associated with the given inode map - * We do basic validation checks on the buffer once it has been - * retrieved from disk. + * The xfs inode contains 2 locks: a multi-reader lock called the + * i_iolock and a multi-reader lock called the i_lock. This routine + * allows either or both of the locks to be obtained. + * + * The 2 locks should always be ordered so that the IO lock is + * obtained first in order to prevent deadlock. + * + * ip -- the inode being locked + * lock_flags -- this parameter indicates the inode's locks + * to be locked. It can be: + * XFS_IOLOCK_SHARED, + * XFS_IOLOCK_EXCL, + * XFS_ILOCK_SHARED, + * XFS_ILOCK_EXCL, + * XFS_IOLOCK_SHARED | XFS_ILOCK_SHARED, + * XFS_IOLOCK_SHARED | XFS_ILOCK_EXCL, + * XFS_IOLOCK_EXCL | XFS_ILOCK_SHARED, + * XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL */ -STATIC int -xfs_imap_to_bp( - xfs_mount_t *mp, - xfs_trans_t *tp, - struct xfs_imap *imap, - xfs_buf_t **bpp, - uint buf_flags, - uint iget_flags) +void +xfs_ilock( + xfs_inode_t *ip, + uint lock_flags) { - int error; - int i; - int ni; - xfs_buf_t *bp; - - error = xfs_trans_read_buf(mp, tp, mp->m_ddev_targp, imap->im_blkno, - (int)imap->im_len, buf_flags, &bp); - if (error) { - if (error != EAGAIN) { - cmn_err(CE_WARN, - "xfs_imap_to_bp: xfs_trans_read_buf()returned " - "an error %d on %s. Returning error.", - error, mp->m_fsname); - } else { - ASSERT(buf_flags & XBF_TRYLOCK); - } - return error; - } + trace_xfs_ilock(ip, lock_flags, _RET_IP_); /* - * Validate the magic number and version of every inode in the buffer - * (if DEBUG kernel) or the first inode in the buffer, otherwise. + * You can't set both SHARED and EXCL for the same lock, + * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED, + * and XFS_ILOCK_EXCL are valid values to set in lock_flags. */ -#ifdef DEBUG - ni = BBTOB(imap->im_len) >> mp->m_sb.sb_inodelog; -#else /* usual case */ - ni = 1; -#endif + ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) != + (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)); + ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) != + (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0); - for (i = 0; i < ni; i++) { - int di_ok; - xfs_dinode_t *dip; - - dip = (xfs_dinode_t *)xfs_buf_offset(bp, - (i << mp->m_sb.sb_inodelog)); - di_ok = be16_to_cpu(dip->di_magic) == XFS_DINODE_MAGIC && - XFS_DINODE_GOOD_VERSION(dip->di_version); - if (unlikely(XFS_TEST_ERROR(!di_ok, mp, - XFS_ERRTAG_ITOBP_INOTOBP, - XFS_RANDOM_ITOBP_INOTOBP))) { - if (iget_flags & XFS_IGET_UNTRUSTED) { - xfs_trans_brelse(tp, bp); - return XFS_ERROR(EINVAL); - } - XFS_CORRUPTION_ERROR("xfs_imap_to_bp", - XFS_ERRLEVEL_HIGH, mp, dip); -#ifdef DEBUG - cmn_err(CE_PANIC, - "Device %s - bad inode magic/vsn " - "daddr %lld #%d (magic=%x)", - XFS_BUFTARG_NAME(mp->m_ddev_targp), - (unsigned long long)imap->im_blkno, i, - be16_to_cpu(dip->di_magic)); -#endif - xfs_trans_brelse(tp, bp); - return XFS_ERROR(EFSCORRUPTED); - } - } - - xfs_inobp_check(mp, bp); - - /* - * Mark the buffer as an inode buffer now that it looks good - */ - XFS_BUF_SET_VTYPE(bp, B_FS_INO); + if (lock_flags & XFS_IOLOCK_EXCL) + mrupdate_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags)); + else if (lock_flags & XFS_IOLOCK_SHARED) + mraccess_nested(&ip->i_iolock, XFS_IOLOCK_DEP(lock_flags)); - *bpp = bp; - return 0; + if (lock_flags & XFS_ILOCK_EXCL) + mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); + else if (lock_flags & XFS_ILOCK_SHARED) + mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags)); } /* - * This routine is called to map an inode number within a file - * system to the buffer containing the on-disk version of the - * inode. It returns a pointer to the buffer containing the - * on-disk inode in the bpp parameter, and in the dip parameter - * it returns a pointer to the on-disk inode within that buffer. + * This is just like xfs_ilock(), except that the caller + * is guaranteed not to sleep. It returns 1 if it gets + * the requested locks and 0 otherwise. If the IO lock is + * obtained but the inode lock cannot be, then the IO lock + * is dropped before returning. * - * If a non-zero error is returned, then the contents of bpp and - * dipp are undefined. - * - * Use xfs_imap() to determine the size and location of the - * buffer to read from disk. + * ip -- the inode being locked + * lock_flags -- this parameter indicates the inode's locks to be + * to be locked. See the comment for xfs_ilock() for a list + * of valid values. */ int -xfs_inotobp( - xfs_mount_t *mp, - xfs_trans_t *tp, - xfs_ino_t ino, - xfs_dinode_t **dipp, - xfs_buf_t **bpp, - int *offset, - uint imap_flags) +xfs_ilock_nowait( + xfs_inode_t *ip, + uint lock_flags) { - struct xfs_imap imap; - xfs_buf_t *bp; - int error; - - imap.im_blkno = 0; - error = xfs_imap(mp, tp, ino, &imap, imap_flags); - if (error) - return error; - - error = xfs_imap_to_bp(mp, tp, &imap, &bp, XBF_LOCK, imap_flags); - if (error) - return error; - - *dipp = (xfs_dinode_t *)xfs_buf_offset(bp, imap.im_boffset); - *bpp = bp; - *offset = imap.im_boffset; + trace_xfs_ilock_nowait(ip, lock_flags, _RET_IP_); + + /* + * You can't set both SHARED and EXCL for the same lock, + * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED, + * and XFS_ILOCK_EXCL are valid values to set in lock_flags. + */ + ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) != + (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)); + ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) != + (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0); + + if (lock_flags & XFS_IOLOCK_EXCL) { + if (!mrtryupdate(&ip->i_iolock)) + goto out; + } else if (lock_flags & XFS_IOLOCK_SHARED) { + if (!mrtryaccess(&ip->i_iolock)) + goto out; + } + if (lock_flags & XFS_ILOCK_EXCL) { + if (!mrtryupdate(&ip->i_lock)) + goto out_undo_iolock; + } else if (lock_flags & XFS_ILOCK_SHARED) { + if (!mrtryaccess(&ip->i_lock)) + goto out_undo_iolock; + } + return 1; + + out_undo_iolock: + if (lock_flags & XFS_IOLOCK_EXCL) + mrunlock_excl(&ip->i_iolock); + else if (lock_flags & XFS_IOLOCK_SHARED) + mrunlock_shared(&ip->i_iolock); + out: return 0; } - /* - * This routine is called to map an inode to the buffer containing - * the on-disk version of the inode. It returns a pointer to the - * buffer containing the on-disk inode in the bpp parameter, and in - * the dip parameter it returns a pointer to the on-disk inode within - * that buffer. + * xfs_iunlock() is used to drop the inode locks acquired with + * xfs_ilock() and xfs_ilock_nowait(). The caller must pass + * in the flags given to xfs_ilock() or xfs_ilock_nowait() so + * that we know which locks to drop. * - * If a non-zero error is returned, then the contents of bpp and - * dipp are undefined. + * ip -- the inode being unlocked + * lock_flags -- this parameter indicates the inode's locks to be + * to be unlocked. See the comment for xfs_ilock() for a list + * of valid values for this parameter. * - * The inode is expected to already been mapped to its buffer and read - * in once, thus we can use the mapping information stored in the inode - * rather than calling xfs_imap(). This allows us to avoid the overhead - * of looking at the inode btree for small block file systems - * (see xfs_imap()). */ -int -xfs_itobp( - xfs_mount_t *mp, - xfs_trans_t *tp, - xfs_inode_t *ip, - xfs_dinode_t **dipp, - xfs_buf_t **bpp, - uint buf_flags) +void +xfs_iunlock( + xfs_inode_t *ip, + uint lock_flags) { - xfs_buf_t *bp; - int error; + /* + * You can't set both SHARED and EXCL for the same lock, + * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED, + * and XFS_ILOCK_EXCL are valid values to set in lock_flags. + */ + ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) != + (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)); + ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) != + (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)); + ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_DEP_MASK)) == 0); + ASSERT(lock_flags != 0); - ASSERT(ip->i_imap.im_blkno != 0); + if (lock_flags & XFS_IOLOCK_EXCL) + mrunlock_excl(&ip->i_iolock); + else if (lock_flags & XFS_IOLOCK_SHARED) + mrunlock_shared(&ip->i_iolock); - error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &bp, buf_flags, 0); - if (error) - return error; - - if (!bp) { - ASSERT(buf_flags & XBF_TRYLOCK); - ASSERT(tp == NULL); - *bpp = NULL; - return EAGAIN; - } + if (lock_flags & XFS_ILOCK_EXCL) + mrunlock_excl(&ip->i_lock); + else if (lock_flags & XFS_ILOCK_SHARED) + mrunlock_shared(&ip->i_lock); - *dipp = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset); - *bpp = bp; - return 0; + trace_xfs_iunlock(ip, lock_flags, _RET_IP_); } /* - * Move inode type and inode format specific information from the - * on-disk inode to the in-core inode. For fifos, devs, and sockets - * this means set if_rdev to the proper value. For files, directories, - * and symlinks this means to bring in the in-line data or extent - * pointers. For a file in B-tree format, only the root is immediately - * brought in-core. The rest will be in-lined in if_extents when it - * is first referenced (see xfs_iread_extents()). + * give up write locks. the i/o lock cannot be held nested + * if it is being demoted. */ -STATIC int -xfs_iformat( +void +xfs_ilock_demote( xfs_inode_t *ip, - xfs_dinode_t *dip) + uint lock_flags) { - xfs_attr_shortform_t *atp; - int size; - int error; - xfs_fsize_t di_size; - ip->i_df.if_ext_max = - XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t); - error = 0; + ASSERT(lock_flags & (XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL)); + ASSERT((lock_flags & ~(XFS_IOLOCK_EXCL|XFS_ILOCK_EXCL)) == 0); - if (unlikely(be32_to_cpu(dip->di_nextents) + - be16_to_cpu(dip->di_anextents) > - be64_to_cpu(dip->di_nblocks))) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt dinode %Lu, extent total = %d, nblocks = %Lu.", - (unsigned long long)ip->i_ino, - (int)(be32_to_cpu(dip->di_nextents) + - be16_to_cpu(dip->di_anextents)), - (unsigned long long) - be64_to_cpu(dip->di_nblocks)); - XFS_CORRUPTION_ERROR("xfs_iformat(1)", XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } + if (lock_flags & XFS_ILOCK_EXCL) + mrdemote(&ip->i_lock); + if (lock_flags & XFS_IOLOCK_EXCL) + mrdemote(&ip->i_iolock); - if (unlikely(dip->di_forkoff > ip->i_mount->m_sb.sb_inodesize)) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt dinode %Lu, forkoff = 0x%x.", - (unsigned long long)ip->i_ino, - dip->di_forkoff); - XFS_CORRUPTION_ERROR("xfs_iformat(2)", XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } + trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_); +} - if (unlikely((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) && - !ip->i_mount->m_rtdev_targp)) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt dinode %Lu, has realtime flag set.", - ip->i_ino); - XFS_CORRUPTION_ERROR("xfs_iformat(realtime)", - XFS_ERRLEVEL_LOW, ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); +#if defined(DEBUG) || defined(XFS_WARN) +int +xfs_isilocked( + xfs_inode_t *ip, + uint lock_flags) +{ + if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) { + if (!(lock_flags & XFS_ILOCK_SHARED)) + return !!ip->i_lock.mr_writer; + return rwsem_is_locked(&ip->i_lock.mr_lock); } - switch (ip->i_d.di_mode & S_IFMT) { - case S_IFIFO: - case S_IFCHR: - case S_IFBLK: - case S_IFSOCK: - if (unlikely(dip->di_format != XFS_DINODE_FMT_DEV)) { - XFS_CORRUPTION_ERROR("xfs_iformat(3)", XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } - ip->i_d.di_size = 0; - ip->i_size = 0; - ip->i_df.if_u2.if_rdev = xfs_dinode_get_rdev(dip); - break; - - case S_IFREG: - case S_IFLNK: - case S_IFDIR: - switch (dip->di_format) { - case XFS_DINODE_FMT_LOCAL: - /* - * no local regular files yet - */ - if (unlikely((be16_to_cpu(dip->di_mode) & S_IFMT) == S_IFREG)) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu " - "(local format for regular file).", - (unsigned long long) ip->i_ino); - XFS_CORRUPTION_ERROR("xfs_iformat(4)", - XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } - - di_size = be64_to_cpu(dip->di_size); - if (unlikely(di_size > XFS_DFORK_DSIZE(dip, ip->i_mount))) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu " - "(bad size %Ld for local inode).", - (unsigned long long) ip->i_ino, - (long long) di_size); - XFS_CORRUPTION_ERROR("xfs_iformat(5)", - XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } - - size = (int)di_size; - error = xfs_iformat_local(ip, dip, XFS_DATA_FORK, size); - break; - case XFS_DINODE_FMT_EXTENTS: - error = xfs_iformat_extents(ip, dip, XFS_DATA_FORK); - break; - case XFS_DINODE_FMT_BTREE: - error = xfs_iformat_btree(ip, dip, XFS_DATA_FORK); - break; - default: - XFS_ERROR_REPORT("xfs_iformat(6)", XFS_ERRLEVEL_LOW, - ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); - } - break; - - default: - XFS_ERROR_REPORT("xfs_iformat(7)", XFS_ERRLEVEL_LOW, ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); + if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) { + if (!(lock_flags & XFS_IOLOCK_SHARED)) + return !!ip->i_iolock.mr_writer; + return rwsem_is_locked(&ip->i_iolock.mr_lock); } - if (error) { - return error; - } - if (!XFS_DFORK_Q(dip)) - return 0; - ASSERT(ip->i_afp == NULL); - ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_SLEEP | KM_NOFS); - ip->i_afp->if_ext_max = - XFS_IFORK_ASIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t); - switch (dip->di_aformat) { - case XFS_DINODE_FMT_LOCAL: - atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip); - size = be16_to_cpu(atp->hdr.totsize); - - if (unlikely(size < sizeof(struct xfs_attr_sf_hdr))) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu " - "(bad attr fork size %Ld).", - (unsigned long long) ip->i_ino, - (long long) size); - XFS_CORRUPTION_ERROR("xfs_iformat(8)", - XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } - error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK, size); - break; - case XFS_DINODE_FMT_EXTENTS: - error = xfs_iformat_extents(ip, dip, XFS_ATTR_FORK); - break; - case XFS_DINODE_FMT_BTREE: - error = xfs_iformat_btree(ip, dip, XFS_ATTR_FORK); - break; - default: - error = XFS_ERROR(EFSCORRUPTED); - break; - } - if (error) { - kmem_zone_free(xfs_ifork_zone, ip->i_afp); - ip->i_afp = NULL; - xfs_idestroy_fork(ip, XFS_DATA_FORK); - } - return error; + ASSERT(0); + return 0; } +#endif + +#ifdef DEBUG +int xfs_locked_n; +int xfs_small_retries; +int xfs_middle_retries; +int xfs_lots_retries; +int xfs_lock_delays; +#endif /* - * The file is in-lined in the on-disk inode. - * If it fits into if_inline_data, then copy - * it there, otherwise allocate a buffer for it - * and copy the data there. Either way, set - * if_data to point at the data. - * If we allocate a buffer for the data, make - * sure that its size is a multiple of 4 and - * record the real size in i_real_bytes. + * Bump the subclass so xfs_lock_inodes() acquires each lock with + * a different value */ -STATIC int -xfs_iformat_local( - xfs_inode_t *ip, - xfs_dinode_t *dip, - int whichfork, - int size) +static inline int +xfs_lock_inumorder(int lock_mode, int subclass) { - xfs_ifork_t *ifp; - int real_size; + if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) + lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_IOLOCK_SHIFT; + if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) + lock_mode |= (subclass + XFS_LOCK_INUMORDER) << XFS_ILOCK_SHIFT; - /* - * If the size is unreasonable, then something - * is wrong and we just bail out rather than crash in - * kmem_alloc() or memcpy() below. - */ - if (unlikely(size > XFS_DFORK_SIZE(dip, ip->i_mount, whichfork))) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu " - "(bad size %d for local fork, size = %d).", - (unsigned long long) ip->i_ino, size, - XFS_DFORK_SIZE(dip, ip->i_mount, whichfork)); - XFS_CORRUPTION_ERROR("xfs_iformat_local", XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } - ifp = XFS_IFORK_PTR(ip, whichfork); - real_size = 0; - if (size == 0) - ifp->if_u1.if_data = NULL; - else if (size <= sizeof(ifp->if_u2.if_inline_data)) - ifp->if_u1.if_data = ifp->if_u2.if_inline_data; - else { - real_size = roundup(size, 4); - ifp->if_u1.if_data = kmem_alloc(real_size, KM_SLEEP | KM_NOFS); - } - ifp->if_bytes = size; - ifp->if_real_bytes = real_size; - if (size) - memcpy(ifp->if_u1.if_data, XFS_DFORK_PTR(dip, whichfork), size); - ifp->if_flags &= ~XFS_IFEXTENTS; - ifp->if_flags |= XFS_IFINLINE; - return 0; + return lock_mode; } /* - * The file consists of a set of extents all - * of which fit into the on-disk inode. - * If there are few enough extents to fit into - * the if_inline_ext, then copy them there. - * Otherwise allocate a buffer for them and copy - * them into it. Either way, set if_extents - * to point at the extents. + * The following routine will lock n inodes in exclusive mode. + * We assume the caller calls us with the inodes in i_ino order. + * + * We need to detect deadlock where an inode that we lock + * is in the AIL and we start waiting for another inode that is locked + * by a thread in a long running transaction (such as truncate). This can + * result in deadlock since the long running trans might need to wait + * for the inode we just locked in order to push the tail and free space + * in the log. */ -STATIC int -xfs_iformat_extents( - xfs_inode_t *ip, - xfs_dinode_t *dip, - int whichfork) +void +xfs_lock_inodes( + xfs_inode_t **ips, + int inodes, + uint lock_mode) { - xfs_bmbt_rec_t *dp; - xfs_ifork_t *ifp; - int nex; - int size; - int i; + int attempts = 0, i, j, try_lock; + xfs_log_item_t *lp; - ifp = XFS_IFORK_PTR(ip, whichfork); - nex = XFS_DFORK_NEXTENTS(dip, whichfork); - size = nex * (uint)sizeof(xfs_bmbt_rec_t); + ASSERT(ips && (inodes >= 2)); /* we need at least two */ - /* - * If the number of extents is unreasonable, then something - * is wrong and we just bail out rather than crash in - * kmem_alloc() or memcpy() below. - */ - if (unlikely(size < 0 || size > XFS_DFORK_SIZE(dip, ip->i_mount, whichfork))) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu ((a)extents = %d).", - (unsigned long long) ip->i_ino, nex); - XFS_CORRUPTION_ERROR("xfs_iformat_extents(1)", XFS_ERRLEVEL_LOW, - ip->i_mount, dip); - return XFS_ERROR(EFSCORRUPTED); - } + try_lock = 0; + i = 0; + +again: + for (; i < inodes; i++) { + ASSERT(ips[i]); + + if (i && (ips[i] == ips[i-1])) /* Already locked */ + continue; - ifp->if_real_bytes = 0; - if (nex == 0) - ifp->if_u1.if_extents = NULL; - else if (nex <= XFS_INLINE_EXTS) - ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext; - else - xfs_iext_add(ifp, 0, nex); - - ifp->if_bytes = size; - if (size) { - dp = (xfs_bmbt_rec_t *) XFS_DFORK_PTR(dip, whichfork); - xfs_validate_extents(ifp, nex, XFS_EXTFMT_INODE(ip)); - for (i = 0; i < nex; i++, dp++) { - xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i); - ep->l0 = get_unaligned_be64(&dp->l0); - ep->l1 = get_unaligned_be64(&dp->l1); + /* + * If try_lock is not set yet, make sure all locked inodes + * are not in the AIL. + * If any are, set try_lock to be used later. + */ + + if (!try_lock) { + for (j = (i - 1); j >= 0 && !try_lock; j--) { + lp = (xfs_log_item_t *)ips[j]->i_itemp; + if (lp && (lp->li_flags & XFS_LI_IN_AIL)) { + try_lock++; + } + } } - XFS_BMAP_TRACE_EXLIST(ip, nex, whichfork); - if (whichfork != XFS_DATA_FORK || - XFS_EXTFMT_INODE(ip) == XFS_EXTFMT_NOSTATE) - if (unlikely(xfs_check_nostate_extents( - ifp, 0, nex))) { - XFS_ERROR_REPORT("xfs_iformat_extents(2)", - XFS_ERRLEVEL_LOW, - ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); + + /* + * If any of the previous locks we have locked is in the AIL, + * we must TRY to get the second and subsequent locks. If + * we can't get any, we must release all we have + * and try again. + */ + + if (try_lock) { + /* try_lock must be 0 if i is 0. */ + /* + * try_lock means we have an inode locked + * that is in the AIL. + */ + ASSERT(i != 0); + if (!xfs_ilock_nowait(ips[i], xfs_lock_inumorder(lock_mode, i))) { + attempts++; + + /* + * Unlock all previous guys and try again. + * xfs_iunlock will try to push the tail + * if the inode is in the AIL. + */ + + for(j = i - 1; j >= 0; j--) { + + /* + * Check to see if we've already + * unlocked this one. + * Not the first one going back, + * and the inode ptr is the same. + */ + if ((j != (i - 1)) && ips[j] == + ips[j+1]) + continue; + + xfs_iunlock(ips[j], lock_mode); + } + + if ((attempts % 5) == 0) { + delay(1); /* Don't just spin the CPU */ +#ifdef DEBUG + xfs_lock_delays++; +#endif } + i = 0; + try_lock = 0; + goto again; + } + } else { + xfs_ilock(ips[i], xfs_lock_inumorder(lock_mode, i)); + } } - ifp->if_flags |= XFS_IFEXTENTS; - return 0; + +#ifdef DEBUG + if (attempts) { + if (attempts < 5) xfs_small_retries++; + else if (attempts < 100) xfs_middle_retries++; + else xfs_lots_retries++; + } else { + xfs_locked_n++; + } +#endif } /* - * The file has too many extents to fit into - * the inode, so they are in B-tree format. - * Allocate a buffer for the root of the B-tree - * and copy the root into it. The i_extents - * field will remain NULL until all of the - * extents are read in (when they are needed). + * xfs_lock_two_inodes() can only be used to lock one type of lock + * at a time - the iolock or the ilock, but not both at once. If + * we lock both at once, lockdep will report false positives saying + * we have violated locking orders. */ -STATIC int -xfs_iformat_btree( - xfs_inode_t *ip, - xfs_dinode_t *dip, - int whichfork) +void +xfs_lock_two_inodes( + xfs_inode_t *ip0, + xfs_inode_t *ip1, + uint lock_mode) { - xfs_bmdr_block_t *dfp; - xfs_ifork_t *ifp; - /* REFERENCED */ - int nrecs; - int size; + xfs_inode_t *temp; + int attempts = 0; + xfs_log_item_t *lp; - ifp = XFS_IFORK_PTR(ip, whichfork); - dfp = (xfs_bmdr_block_t *)XFS_DFORK_PTR(dip, whichfork); - size = XFS_BMAP_BROOT_SPACE(dfp); - nrecs = be16_to_cpu(dfp->bb_numrecs); + if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) + ASSERT((lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) == 0); + ASSERT(ip0->i_ino != ip1->i_ino); - /* - * blow out if -- fork has less extents than can fit in - * fork (fork shouldn't be a btree format), root btree - * block has more records than can fit into the fork, - * or the number of extents is greater than the number of - * blocks. - */ - if (unlikely(XFS_IFORK_NEXTENTS(ip, whichfork) <= ifp->if_ext_max - || XFS_BMDR_SPACE_CALC(nrecs) > - XFS_DFORK_SIZE(dip, ip->i_mount, whichfork) - || XFS_IFORK_NEXTENTS(ip, whichfork) > ip->i_d.di_nblocks)) { - xfs_fs_repair_cmn_err(CE_WARN, ip->i_mount, - "corrupt inode %Lu (btree).", - (unsigned long long) ip->i_ino); - XFS_ERROR_REPORT("xfs_iformat_btree", XFS_ERRLEVEL_LOW, - ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); + if (ip0->i_ino > ip1->i_ino) { + temp = ip0; + ip0 = ip1; + ip1 = temp; } - ifp->if_broot_bytes = size; - ifp->if_broot = kmem_alloc(size, KM_SLEEP | KM_NOFS); - ASSERT(ifp->if_broot != NULL); + again: + xfs_ilock(ip0, xfs_lock_inumorder(lock_mode, 0)); + /* - * Copy and convert from the on-disk structure - * to the in-memory structure. + * If the first lock we have locked is in the AIL, we must TRY to get + * the second lock. If we can't get it, we must release the first one + * and try again. */ - xfs_bmdr_to_bmbt(ip->i_mount, dfp, - XFS_DFORK_SIZE(dip, ip->i_mount, whichfork), - ifp->if_broot, size); - ifp->if_flags &= ~XFS_IFEXTENTS; - ifp->if_flags |= XFS_IFBROOT; - - return 0; + lp = (xfs_log_item_t *)ip0->i_itemp; + if (lp && (lp->li_flags & XFS_LI_IN_AIL)) { + if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(lock_mode, 1))) { + xfs_iunlock(ip0, lock_mode); + if ((++attempts % 5) == 0) + delay(1); /* Don't just spin the CPU */ + goto again; + } + } else { + xfs_ilock(ip1, xfs_lock_inumorder(lock_mode, 1)); + } } -STATIC void -xfs_dinode_from_disk( - xfs_icdinode_t *to, - xfs_dinode_t *from) -{ - to->di_magic = be16_to_cpu(from->di_magic); - to->di_mode = be16_to_cpu(from->di_mode); - to->di_version = from ->di_version; - to->di_format = from->di_format; - to->di_onlink = be16_to_cpu(from->di_onlink); - to->di_uid = be32_to_cpu(from->di_uid); - to->di_gid = be32_to_cpu(from->di_gid); - to->di_nlink = be32_to_cpu(from->di_nlink); - to->di_projid_lo = be16_to_cpu(from->di_projid_lo); - to->di_projid_hi = be16_to_cpu(from->di_projid_hi); - memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad)); - to->di_flushiter = be16_to_cpu(from->di_flushiter); - to->di_atime.t_sec = be32_to_cpu(from->di_atime.t_sec); - to->di_atime.t_nsec = be32_to_cpu(from->di_atime.t_nsec); - to->di_mtime.t_sec = be32_to_cpu(from->di_mtime.t_sec); - to->di_mtime.t_nsec = be32_to_cpu(from->di_mtime.t_nsec); - to->di_ctime.t_sec = be32_to_cpu(from->di_ctime.t_sec); - to->di_ctime.t_nsec = be32_to_cpu(from->di_ctime.t_nsec); - to->di_size = be64_to_cpu(from->di_size); - to->di_nblocks = be64_to_cpu(from->di_nblocks); - to->di_extsize = be32_to_cpu(from->di_extsize); - to->di_nextents = be32_to_cpu(from->di_nextents); - to->di_anextents = be16_to_cpu(from->di_anextents); - to->di_forkoff = from->di_forkoff; - to->di_aformat = from->di_aformat; - to->di_dmevmask = be32_to_cpu(from->di_dmevmask); - 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); -} void -xfs_dinode_to_disk( - xfs_dinode_t *to, - xfs_icdinode_t *from) +__xfs_iflock( + struct xfs_inode *ip) { - to->di_magic = cpu_to_be16(from->di_magic); - to->di_mode = cpu_to_be16(from->di_mode); - to->di_version = from ->di_version; - to->di_format = from->di_format; - to->di_onlink = cpu_to_be16(from->di_onlink); - to->di_uid = cpu_to_be32(from->di_uid); - to->di_gid = cpu_to_be32(from->di_gid); - to->di_nlink = cpu_to_be32(from->di_nlink); - to->di_projid_lo = cpu_to_be16(from->di_projid_lo); - to->di_projid_hi = cpu_to_be16(from->di_projid_hi); - memcpy(to->di_pad, from->di_pad, sizeof(to->di_pad)); - to->di_flushiter = cpu_to_be16(from->di_flushiter); - to->di_atime.t_sec = cpu_to_be32(from->di_atime.t_sec); - to->di_atime.t_nsec = cpu_to_be32(from->di_atime.t_nsec); - to->di_mtime.t_sec = cpu_to_be32(from->di_mtime.t_sec); - to->di_mtime.t_nsec = cpu_to_be32(from->di_mtime.t_nsec); - to->di_ctime.t_sec = cpu_to_be32(from->di_ctime.t_sec); - to->di_ctime.t_nsec = cpu_to_be32(from->di_ctime.t_nsec); - to->di_size = cpu_to_be64(from->di_size); - to->di_nblocks = cpu_to_be64(from->di_nblocks); - to->di_extsize = cpu_to_be32(from->di_extsize); - to->di_nextents = cpu_to_be32(from->di_nextents); - to->di_anextents = cpu_to_be16(from->di_anextents); - to->di_forkoff = from->di_forkoff; - to->di_aformat = from->di_aformat; - to->di_dmevmask = cpu_to_be32(from->di_dmevmask); - 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); + wait_queue_head_t *wq = bit_waitqueue(&ip->i_flags, __XFS_IFLOCK_BIT); + DEFINE_WAIT_BIT(wait, &ip->i_flags, __XFS_IFLOCK_BIT); + + do { + prepare_to_wait_exclusive(wq, &wait.wait, TASK_UNINTERRUPTIBLE); + if (xfs_isiflocked(ip)) + io_schedule(); + } while (!xfs_iflock_nowait(ip)); + + finish_wait(wq, &wait.wait); } STATIC uint @@ -778,191 +564,65 @@ xfs_dic2xflags( } /* - * Read the disk inode attributes into the in-core inode structure. + * Lookups up an inode from "name". If ci_name is not NULL, then a CI match + * is allowed, otherwise it has to be an exact match. If a CI match is found, + * ci_name->name will point to a the actual name (caller must free) or + * will be set to NULL if an exact match is found. */ int -xfs_iread( - xfs_mount_t *mp, - xfs_trans_t *tp, - xfs_inode_t *ip, - uint iget_flags) +xfs_lookup( + xfs_inode_t *dp, + struct xfs_name *name, + xfs_inode_t **ipp, + struct xfs_name *ci_name) { - xfs_buf_t *bp; - xfs_dinode_t *dip; - int error; - - /* - * Fill in the location information in the in-core inode. - */ - error = xfs_imap(mp, tp, ip->i_ino, &ip->i_imap, iget_flags); - if (error) - return error; + xfs_ino_t inum; + int error; + uint lock_mode; - /* - * Get pointers to the on-disk inode and the buffer containing it. - */ - error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &bp, - XBF_LOCK, iget_flags); - if (error) - return error; - dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset); + trace_xfs_lookup(dp, name); - /* - * If we got something that isn't an inode it means someone - * (nfs or dmi) has a stale handle. - */ - if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC) { -#ifdef DEBUG - xfs_fs_cmn_err(CE_ALERT, mp, "xfs_iread: " - "dip->di_magic (0x%x) != " - "XFS_DINODE_MAGIC (0x%x)", - be16_to_cpu(dip->di_magic), - XFS_DINODE_MAGIC); -#endif /* DEBUG */ - error = XFS_ERROR(EINVAL); - goto out_brelse; - } + if (XFS_FORCED_SHUTDOWN(dp->i_mount)) + return XFS_ERROR(EIO); - /* - * If the on-disk inode is already linked to a directory - * entry, copy all of the inode into the in-core inode. - * xfs_iformat() handles copying in the inode format - * specific information. - * Otherwise, just get the truly permanent information. - */ - if (dip->di_mode) { - xfs_dinode_from_disk(&ip->i_d, dip); - error = xfs_iformat(ip, dip); - if (error) { -#ifdef DEBUG - xfs_fs_cmn_err(CE_ALERT, mp, "xfs_iread: " - "xfs_iformat() returned error %d", - error); -#endif /* DEBUG */ - goto out_brelse; - } - } else { - 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); - /* - * Make sure to pull in the mode here as well in - * case the inode is released without being used. - * This ensures that xfs_inactive() will see that - * the inode is already free and not try to mess - * with the uninitialized part of it. - */ - ip->i_d.di_mode = 0; - /* - * Initialize the per-fork minima and maxima for a new - * inode here. xfs_iformat will do it for old inodes. - */ - ip->i_df.if_ext_max = - XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t); - } + lock_mode = xfs_ilock_data_map_shared(dp); + error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name); + xfs_iunlock(dp, lock_mode); - /* - * The inode format changed when we moved the link count and - * made it 32 bits long. If this is an old format inode, - * convert it in memory to look like a new one. If it gets - * flushed to disk we will convert back before flushing or - * logging it. We zero out the new projid field and the old link - * count field. We'll handle clearing the pad field (the remains - * of the old uuid field) when we actually convert the inode to - * the new format. We don't change the version number so that we - * can distinguish this from a real new format inode. - */ - if (ip->i_d.di_version == 1) { - ip->i_d.di_nlink = ip->i_d.di_onlink; - ip->i_d.di_onlink = 0; - xfs_set_projid(ip, 0); - } + if (error) + goto out; - ip->i_delayed_blks = 0; - ip->i_size = ip->i_d.di_size; + error = xfs_iget(dp->i_mount, NULL, inum, 0, 0, ipp); + if (error) + goto out_free_name; - /* - * Mark the buffer containing the inode as something to keep - * around for a while. This helps to keep recently accessed - * meta-data in-core longer. - */ - XFS_BUF_SET_REF(bp, XFS_INO_REF); + return 0; - /* - * Use xfs_trans_brelse() to release the buffer containing the - * on-disk inode, because it was acquired with xfs_trans_read_buf() - * in xfs_itobp() above. If tp is NULL, this is just a normal - * brelse(). If we're within a transaction, then xfs_trans_brelse() - * will only release the buffer if it is not dirty within the - * transaction. It will be OK to release the buffer in this case, - * because inodes on disk are never destroyed and we will be - * locking the new in-core inode before putting it in the hash - * table where other processes can find it. Thus we don't have - * to worry about the inode being changed just because we released - * the buffer. - */ - out_brelse: - xfs_trans_brelse(tp, bp); +out_free_name: + if (ci_name) + kmem_free(ci_name->name); +out: + *ipp = NULL; return error; } /* - * Read in extents from a btree-format inode. - * Allocate and fill in if_extents. Real work is done in xfs_bmap.c. - */ -int -xfs_iread_extents( - xfs_trans_t *tp, - xfs_inode_t *ip, - int whichfork) -{ - int error; - xfs_ifork_t *ifp; - xfs_extnum_t nextents; - - if (unlikely(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)) { - XFS_ERROR_REPORT("xfs_iread_extents", XFS_ERRLEVEL_LOW, - ip->i_mount); - return XFS_ERROR(EFSCORRUPTED); - } - nextents = XFS_IFORK_NEXTENTS(ip, whichfork); - ifp = XFS_IFORK_PTR(ip, whichfork); - - /* - * We know that the size is valid (it's checked in iformat_btree) - */ - ifp->if_lastex = NULLEXTNUM; - ifp->if_bytes = ifp->if_real_bytes = 0; - ifp->if_flags |= XFS_IFEXTENTS; - xfs_iext_add(ifp, 0, nextents); - error = xfs_bmap_read_extents(tp, ip, whichfork); - if (error) { - xfs_iext_destroy(ifp); - ifp->if_flags &= ~XFS_IFEXTENTS; - return error; - } - xfs_validate_extents(ifp, nextents, XFS_EXTFMT_INODE(ip)); - return 0; -} - -/* * Allocate an inode on disk and return a copy of its in-core version. * The in-core inode is locked exclusively. Set mode, nlink, and rdev * appropriately within the inode. The uid and gid for the inode are * set according to the contents of the given cred structure. * * Use xfs_dialloc() to allocate the on-disk inode. If xfs_dialloc() - * has a free inode available, call xfs_iget() - * to obtain the in-core version of the allocated inode. Finally, - * fill in the inode and log its initial contents. In this case, - * ialloc_context would be set to NULL and call_again set to false. + * has a free inode available, call xfs_iget() to obtain the in-core + * version of the allocated inode. Finally, fill in the inode and + * log its initial contents. In this case, ialloc_context would be + * set to NULL. * - * If xfs_dialloc() does not have an available inode, - * it will replenish its supply by doing an allocation. Since we can - * only do one allocation within a transaction without deadlocks, we - * must commit the current transaction before returning the inode itself. - * In this case, therefore, we will set call_again to true and return. + * If xfs_dialloc() does not have an available inode, it will replenish + * its supply by doing an allocation. Since we can only do one + * allocation within a transaction without deadlocks, we must commit + * the current transaction before returning the inode itself. + * In this case, therefore, we will set ialloc_context and return. * The caller should then commit the current transaction, start a new * transaction, and call xfs_ialloc() again to actually get the inode. * @@ -981,31 +641,30 @@ int xfs_ialloc( xfs_trans_t *tp, xfs_inode_t *pip, - mode_t mode, + umode_t mode, xfs_nlink_t nlink, xfs_dev_t rdev, prid_t prid, int okalloc, xfs_buf_t **ialloc_context, - boolean_t *call_again, xfs_inode_t **ipp) { + struct xfs_mount *mp = tp->t_mountp; xfs_ino_t ino; xfs_inode_t *ip; uint flags; int error; timespec_t tv; - int filestreams = 0; /* * Call the space management code to pick * the on-disk inode to be allocated. */ error = xfs_dialloc(tp, pip ? pip->i_ino : 0, mode, okalloc, - ialloc_context, call_again, &ino); + ialloc_context, &ino); if (error) return error; - if (*call_again || ino == NULLFSINO) { + if (*ialloc_context || ino == NULLFSINO) { *ipp = NULL; return 0; } @@ -1016,45 +675,32 @@ xfs_ialloc( * This is because we're setting fields here we need * to prevent others from looking at until we're done. */ - error = xfs_trans_iget(tp->t_mountp, tp, ino, - XFS_IGET_CREATE, XFS_ILOCK_EXCL, &ip); + error = xfs_iget(mp, tp, ino, XFS_IGET_CREATE, + XFS_ILOCK_EXCL, &ip); if (error) return error; ASSERT(ip != NULL); - ip->i_d.di_mode = (__uint16_t)mode; + /* + * 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; ASSERT(ip->i_d.di_nlink == nlink); - ip->i_d.di_uid = current_fsuid(); - ip->i_d.di_gid = current_fsgid(); + ip->i_d.di_uid = xfs_kuid_to_uid(current_fsuid()); + ip->i_d.di_gid = xfs_kgid_to_gid(current_fsgid()); 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(&tp->t_mountp->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) && (mode & S_IFMT) == S_IFDIR) { + if ((pip->i_d.di_mode & S_ISGID) && S_ISDIR(mode)) { ip->i_d.di_mode |= S_ISGID; } } @@ -1066,12 +712,11 @@ xfs_ialloc( */ if ((irix_sgid_inherit) && (ip->i_d.di_mode & S_ISGID) && - (!in_group_p((gid_t)ip->i_d.di_gid))) { + (!in_group_p(xfs_gid_to_kgid(ip->i_d.di_gid)))) { ip->i_d.di_mode &= ~S_ISGID; } ip->i_d.di_size = 0; - ip->i_size = 0; ip->i_d.di_nextents = 0; ASSERT(ip->i_d.di_nblocks == 0); @@ -1088,6 +733,19 @@ xfs_ialloc( 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: @@ -1100,25 +758,18 @@ 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; - if ((mode & S_IFMT) == S_IFDIR) { + if (S_ISDIR(mode)) { if (pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT) di_flags |= XFS_DIFLAG_RTINHERIT; if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) { di_flags |= XFS_DIFLAG_EXTSZINHERIT; ip->i_d.di_extsize = pip->i_d.di_extsize; } - } else if ((mode & S_IFMT) == S_IFREG) { + } else if (S_ISREG(mode)) { if (pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT) di_flags |= XFS_DIFLAG_REALTIME; if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) { @@ -1166,228 +817,662 @@ xfs_ialloc( /* * Log the new values stuffed into the inode. */ + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); xfs_trans_log_inode(tp, ip, flags); /* 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; +} + +/* + * Allocates a new inode from disk and return a pointer to the + * incore copy. This routine will internally commit the current + * transaction and allocate a new one if the Space Manager needed + * to do an allocation to replenish the inode free-list. + * + * This routine is designed to be called from xfs_create and + * xfs_create_dir. + * + */ +int +xfs_dir_ialloc( + xfs_trans_t **tpp, /* input: current transaction; + output: may be a new transaction. */ + xfs_inode_t *dp, /* directory within whose allocate + the inode. */ + umode_t mode, + xfs_nlink_t nlink, + xfs_dev_t rdev, + prid_t prid, /* project id */ + int okalloc, /* ok to allocate new space */ + xfs_inode_t **ipp, /* pointer to inode; it will be + locked. */ + int *committed) + +{ + xfs_trans_t *tp; + xfs_trans_t *ntp; + xfs_inode_t *ip; + xfs_buf_t *ialloc_context = NULL; + int code; + void *dqinfo; + uint tflags; + + tp = *tpp; + ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES); + + /* + * xfs_ialloc will return a pointer to an incore inode if + * the Space Manager has an available inode on the free + * list. Otherwise, it will do an allocation and replenish + * the freelist. Since we can only do one allocation per + * transaction without deadlocks, we will need to commit the + * current transaction and start a new one. We will then + * need to call xfs_ialloc again to get the inode. + * + * If xfs_ialloc did an allocation to replenish the freelist, + * it returns the bp containing the head of the freelist as + * ialloc_context. We will hold a lock on it across the + * transaction commit so that no other process can steal + * the inode(s) that we've just allocated. + */ + code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid, okalloc, + &ialloc_context, &ip); + + /* + * Return an error if we were unable to allocate a new inode. + * This should only happen if we run out of space on disk or + * encounter a disk error. + */ + if (code) { + *ipp = NULL; + return code; + } + if (!ialloc_context && !ip) { + *ipp = NULL; + return XFS_ERROR(ENOSPC); + } + + /* + * If the AGI buffer is non-NULL, then we were unable to get an + * inode in one operation. We need to commit the current + * transaction and call xfs_ialloc() again. It is guaranteed + * to succeed the second time. + */ + if (ialloc_context) { + struct xfs_trans_res tres; + + /* + * Normally, xfs_trans_commit releases all the locks. + * We call bhold to hang on to the ialloc_context across + * the commit. Holding this buffer prevents any other + * processes from doing any allocations in this + * allocation group. + */ + xfs_trans_bhold(tp, ialloc_context); + /* + * Save the log reservation so we can use + * them in the next transaction. + */ + tres.tr_logres = xfs_trans_get_log_res(tp); + tres.tr_logcount = xfs_trans_get_log_count(tp); + + /* + * We want the quota changes to be associated with the next + * transaction, NOT this one. So, detach the dqinfo from this + * and attach it to the next transaction. + */ + dqinfo = NULL; + tflags = 0; + if (tp->t_dqinfo) { + dqinfo = (void *)tp->t_dqinfo; + tp->t_dqinfo = NULL; + tflags = tp->t_flags & XFS_TRANS_DQ_DIRTY; + tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY); + } + + ntp = xfs_trans_dup(tp); + code = xfs_trans_commit(tp, 0); + tp = ntp; + if (committed != NULL) { + *committed = 1; + } + /* + * If we get an error during the commit processing, + * release the buffer that is still held and return + * to the caller. + */ + if (code) { + xfs_buf_relse(ialloc_context); + if (dqinfo) { + tp->t_dqinfo = dqinfo; + xfs_trans_free_dqinfo(tp); + } + *tpp = ntp; + *ipp = NULL; + return code; + } + + /* + * transaction commit worked ok so we can drop the extra ticket + * reference that we gained in xfs_trans_dup() + */ + xfs_log_ticket_put(tp->t_ticket); + tres.tr_logflags = XFS_TRANS_PERM_LOG_RES; + code = xfs_trans_reserve(tp, &tres, 0, 0); + + /* + * Re-attach the quota info that we detached from prev trx. + */ + if (dqinfo) { + tp->t_dqinfo = dqinfo; + tp->t_flags |= tflags; + } + + if (code) { + xfs_buf_relse(ialloc_context); + *tpp = ntp; + *ipp = NULL; + return code; + } + xfs_trans_bjoin(tp, ialloc_context); + + /* + * Call ialloc again. Since we've locked out all + * other allocations in this allocation group, + * this call should always succeed. + */ + code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid, + okalloc, &ialloc_context, &ip); + + /* + * If we get an error at this point, return to the caller + * so that the current transaction can be aborted. + */ + if (code) { + *tpp = tp; + *ipp = NULL; + return code; + } + ASSERT(!ialloc_context && ip); + + } else { + if (committed != NULL) + *committed = 0; } *ipp = ip; + *tpp = tp; + return 0; } /* - * Check to make sure that there are no blocks allocated to the - * file beyond the size of the file. We don't check this for - * files with fixed size extents or real time extents, but we - * at least do it for regular files. + * Decrement the link count on an inode & log the change. + * If this causes the link count to go to zero, initiate the + * logging activity required to truncate a file. */ -#ifdef DEBUG -void -xfs_isize_check( - xfs_mount_t *mp, - xfs_inode_t *ip, - xfs_fsize_t isize) +int /* error */ +xfs_droplink( + xfs_trans_t *tp, + xfs_inode_t *ip) { - xfs_fileoff_t map_first; - int nimaps; - xfs_bmbt_irec_t imaps[2]; + int error; - if ((ip->i_d.di_mode & S_IFMT) != S_IFREG) - return; - - if (XFS_IS_REALTIME_INODE(ip)) - return; + xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG); - if (ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE) - return; + ASSERT (ip->i_d.di_nlink > 0); + ip->i_d.di_nlink--; + drop_nlink(VFS_I(ip)); + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - nimaps = 2; - map_first = XFS_B_TO_FSB(mp, (xfs_ufsize_t)isize); - /* - * The filesystem could be shutting down, so bmapi may return - * an error. - */ - if (xfs_bmapi(NULL, ip, map_first, - (XFS_B_TO_FSB(mp, - (xfs_ufsize_t)XFS_MAXIOFFSET(mp)) - - map_first), - XFS_BMAPI_ENTIRE, NULL, 0, imaps, &nimaps, - NULL)) - return; - ASSERT(nimaps == 1); - ASSERT(imaps[0].br_startblock == HOLESTARTBLOCK); + error = 0; + if (ip->i_d.di_nlink == 0) { + /* + * We're dropping the last link to this file. + * Move the on-disk inode to the AGI unlinked list. + * From xfs_inactive() we will pull the inode from + * the list and free it. + */ + error = xfs_iunlink(tp, ip); + } + return error; } -#endif /* DEBUG */ /* - * Calculate the last possible buffered byte in a file. This must - * include data that was buffered beyond the EOF by the write code. - * This also needs to deal with overflowing the xfs_fsize_t type - * which can happen for sizes near the limit. - * - * We also need to take into account any blocks beyond the EOF. It - * may be the case that they were buffered by a write which failed. - * In that case the pages will still be in memory, but the inode size - * will never have been updated. + * Increment the link count on an inode & log the change. */ -STATIC xfs_fsize_t -xfs_file_last_byte( - xfs_inode_t *ip) +int +xfs_bumplink( + xfs_trans_t *tp, + xfs_inode_t *ip) { - xfs_mount_t *mp; - xfs_fsize_t last_byte; - xfs_fileoff_t last_block; - xfs_fileoff_t size_last_block; - int error; + xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG); - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)); + 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)); + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + return 0; +} + +int +xfs_create( + xfs_inode_t *dp, + struct xfs_name *name, + umode_t mode, + xfs_dev_t rdev, + xfs_inode_t **ipp) +{ + int is_dir = S_ISDIR(mode); + struct xfs_mount *mp = dp->i_mount; + struct xfs_inode *ip = NULL; + struct xfs_trans *tp = NULL; + int error; + xfs_bmap_free_t free_list; + xfs_fsblock_t first_block; + bool unlock_dp_on_error = false; + uint cancel_flags; + int committed; + 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; + + trace_xfs_create(dp, name); + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + prid = xfs_get_initial_prid(dp); - mp = ip->i_mount; /* - * Only check for blocks beyond the EOF if the extents have - * been read in. This eliminates the need for the inode lock, - * and it also saves us from looking when it really isn't - * necessary. + * Make sure that we have allocated dquot(s) on disk. */ - if (ip->i_df.if_flags & XFS_IFEXTENTS) { - xfs_ilock(ip, XFS_ILOCK_SHARED); - error = xfs_bmap_last_offset(NULL, ip, &last_block, - XFS_DATA_FORK); - xfs_iunlock(ip, XFS_ILOCK_SHARED); - if (error) { - last_block = 0; - } + 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; + + if (is_dir) { + rdev = 0; + resblks = XFS_MKDIR_SPACE_RES(mp, name->len); + tres.tr_logres = M_RES(mp)->tr_mkdir.tr_logres; + tres.tr_logcount = XFS_MKDIR_LOG_COUNT; + tp = xfs_trans_alloc(mp, XFS_TRANS_MKDIR); } else { - last_block = 0; + resblks = XFS_CREATE_SPACE_RES(mp, name->len); + tres.tr_logres = M_RES(mp)->tr_create.tr_logres; + tres.tr_logcount = XFS_CREATE_LOG_COUNT; + tp = xfs_trans_alloc(mp, XFS_TRANS_CREATE); } - size_last_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)ip->i_size); - last_block = XFS_FILEOFF_MAX(last_block, size_last_block); - last_byte = XFS_FSB_TO_B(mp, last_block); - if (last_byte < 0) { - return XFS_MAXIOFFSET(mp); + cancel_flags = XFS_TRANS_RELEASE_LOG_RES; + + /* + * Initially assume that the file does not exist and + * reserve the resources for that case. If that is not + * the case we'll drop the one we have and get a more + * appropriate transaction later. + */ + tres.tr_logflags = XFS_TRANS_PERM_LOG_RES; + error = xfs_trans_reserve(tp, &tres, resblks, 0); + if (error == ENOSPC) { + /* flush outstanding delalloc blocks and retry */ + xfs_flush_inodes(mp); + error = xfs_trans_reserve(tp, &tres, resblks, 0); } - last_byte += (1 << mp->m_writeio_log); - if (last_byte < 0) { - return XFS_MAXIOFFSET(mp); + if (error == ENOSPC) { + /* No space at all so try a "no-allocation" reservation */ + resblks = 0; + error = xfs_trans_reserve(tp, &tres, 0, 0); } - return last_byte; + if (error) { + cancel_flags = 0; + goto out_trans_cancel; + } + + xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT); + unlock_dp_on_error = true; + + xfs_bmap_init(&free_list, &first_block); + + /* + * Reserve disk quota and the inode. + */ + error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp, + pdqp, resblks, 1, 0); + if (error) + goto out_trans_cancel; + + error = xfs_dir_canenter(tp, dp, name, resblks); + if (error) + goto out_trans_cancel; + + /* + * A newly created regular or special file just has one directory + * entry pointing to them, but a directory also the "." entry + * pointing to itself. + */ + error = xfs_dir_ialloc(&tp, dp, mode, is_dir ? 2 : 1, rdev, + prid, resblks > 0, &ip, &committed); + if (error) { + if (error == ENOSPC) + goto out_trans_cancel; + goto out_trans_abort; + } + + /* + * Now we join the directory inode to the transaction. We do not do it + * earlier because xfs_dir_ialloc might commit the previous transaction + * (and release all the locks). An error from here on will result in + * the transaction cancel unlocking dp so don't do it explicitly in the + * error path. + */ + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); + unlock_dp_on_error = false; + + error = xfs_dir_createname(tp, dp, name, ip->i_ino, + &first_block, &free_list, resblks ? + resblks - XFS_IALLOC_SPACE_RES(mp) : 0); + if (error) { + ASSERT(error != ENOSPC); + goto out_trans_abort; + } + xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE); + + if (is_dir) { + error = xfs_dir_init(tp, ip, dp); + if (error) + goto out_bmap_cancel; + + error = xfs_bumplink(tp, dp); + if (error) + goto out_bmap_cancel; + } + + /* + * If this is a synchronous mount, make sure that the + * create transaction goes to disk before returning to + * the user. + */ + if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) + 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); + + error = xfs_bmap_finish(&tp, &free_list, &committed); + if (error) + goto out_bmap_cancel; + + 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_bmap_cancel: + xfs_bmap_cancel(&free_list); + 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); + + if (unlock_dp_on_error) + xfs_iunlock(dp, XFS_ILOCK_EXCL); + return error; } -/* - * Start the truncation of the file to new_size. The new size - * must be smaller than the current size. This routine will - * clear the buffer and page caches of file data in the removed - * range, and xfs_itruncate_finish() will remove the underlying - * disk blocks. - * - * The inode must have its I/O lock locked EXCLUSIVELY, and it - * must NOT have the inode lock held at all. This is because we're - * calling into the buffer/page cache code and we can't hold the - * inode lock when we do so. - * - * We need to wait for any direct I/Os in flight to complete before we - * proceed with the truncate. This is needed to prevent the extents - * being read or written by the direct I/Os from being removed while the - * I/O is in flight as there is no other method of synchronising - * direct I/O with the truncate operation. Also, because we hold - * the IOLOCK in exclusive mode, we prevent new direct I/Os from being - * started until the truncate completes and drops the lock. Essentially, - * the xfs_ioend_wait() call forms an I/O barrier that provides strict - * ordering between direct I/Os and the truncate operation. - * - * The flags parameter can have either the value XFS_ITRUNC_DEFINITE - * or XFS_ITRUNC_MAYBE. The XFS_ITRUNC_MAYBE value should be used - * in the case that the caller is locking things out of order and - * may not be able to call xfs_itruncate_finish() with the inode lock - * held without dropping the I/O lock. If the caller must drop the - * I/O lock before calling xfs_itruncate_finish(), then xfs_itruncate_start() - * must be called again with all the same restrictions as the initial - * call. - */ int -xfs_itruncate_start( - xfs_inode_t *ip, - uint flags, - xfs_fsize_t new_size) +xfs_create_tmpfile( + struct xfs_inode *dp, + struct dentry *dentry, + umode_t mode, + struct xfs_inode **ipp) { - xfs_fsize_t last_byte; - xfs_off_t toss_start; - xfs_mount_t *mp; - int error = 0; + 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); - ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL)); - ASSERT((new_size == 0) || (new_size <= ip->i_size)); - ASSERT((flags == XFS_ITRUNC_DEFINITE) || - (flags == XFS_ITRUNC_MAYBE)); + prid = xfs_get_initial_prid(dp); - mp = ip->i_mount; + /* + * 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; + } - /* wait for the completion of any pending DIOs */ - if (new_size == 0 || new_size < ip->i_size) - xfs_ioend_wait(ip); + if (mp->m_flags & XFS_MOUNT_WSYNC) + xfs_trans_set_sync(tp); /* - * Call toss_pages or flushinval_pages to get rid of pages - * overlapping the region being removed. We have to use - * the less efficient flushinval_pages in the case that the - * caller may not be able to finish the truncate without - * dropping the inode's I/O lock. Make sure - * to catch any pages brought in by buffers overlapping - * the EOF by searching out beyond the isize by our - * block size. We round new_size up to a block boundary - * so that we don't toss things on the same block as - * new_size but before it. - * - * Before calling toss_page or flushinval_pages, make sure to - * call remapf() over the same region if the file is mapped. - * This frees up mapped file references to the pages in the - * given range and for the flushinval_pages case it ensures - * that we get the latest mapped changes flushed out. + * 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. */ - toss_start = XFS_B_TO_FSB(mp, (xfs_ufsize_t)new_size); - toss_start = XFS_FSB_TO_B(mp, toss_start); - if (toss_start < 0) { - /* - * The place to start tossing is beyond our maximum - * file size, so there is no way that the data extended - * out there. - */ - return 0; + 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, + struct xfs_name *target_name) +{ + xfs_mount_t *mp = tdp->i_mount; + xfs_trans_t *tp; + int error; + xfs_bmap_free_t free_list; + xfs_fsblock_t first_block; + int cancel_flags; + int committed; + int resblks; + + trace_xfs_link(tdp, target_name); + + ASSERT(!S_ISDIR(sip->i_d.di_mode)); + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); + + error = xfs_qm_dqattach(sip, 0); + if (error) + goto std_return; + + error = xfs_qm_dqattach(tdp, 0); + if (error) + goto std_return; + + tp = xfs_trans_alloc(mp, XFS_TRANS_LINK); + cancel_flags = XFS_TRANS_RELEASE_LOG_RES; + resblks = XFS_LINK_SPACE_RES(mp, target_name->len); + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, resblks, 0); + if (error == ENOSPC) { + resblks = 0; + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_link, 0, 0); } - last_byte = xfs_file_last_byte(ip); - trace_xfs_itruncate_start(ip, flags, new_size, toss_start, last_byte); - if (last_byte > toss_start) { - if (flags & XFS_ITRUNC_DEFINITE) { - xfs_tosspages(ip, toss_start, - -1, FI_REMAPF_LOCKED); - } else { - error = xfs_flushinval_pages(ip, toss_start, - -1, FI_REMAPF_LOCKED); - } + if (error) { + cancel_flags = 0; + goto error_return; } -#ifdef DEBUG - if (new_size == 0) { - ASSERT(VN_CACHED(VFS_I(ip)) == 0); + xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL); + + xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL); + + /* + * If we are using project inheritance, we only allow hard link + * creation in our tree when the project IDs are the same; else + * the tree quota mechanism could be circumvented. + */ + if (unlikely((tdp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) && + (xfs_get_projid(tdp) != xfs_get_projid(sip)))) { + error = XFS_ERROR(EXDEV); + goto error_return; } -#endif + + error = xfs_dir_canenter(tp, tdp, target_name, resblks); + if (error) + goto error_return; + + 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) + goto abort_return; + xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE); + + error = xfs_bumplink(tp, sip); + if (error) + goto abort_return; + + /* + * If this is a synchronous mount, make sure that the + * link transaction goes to disk before returning to + * the user. + */ + if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) { + xfs_trans_set_sync(tp); + } + + error = xfs_bmap_finish (&tp, &free_list, &committed); + if (error) { + xfs_bmap_cancel(&free_list); + goto abort_return; + } + + return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + + abort_return: + cancel_flags |= XFS_TRANS_ABORT; + error_return: + xfs_trans_cancel(tp, cancel_flags); + std_return: return error; } /* - * Shrink the file to the given new_size. The new size must be smaller than - * the current size. This will free up the underlying blocks in the removed - * range after a call to xfs_itruncate_start() or xfs_atruncate_start(). + * Free up the underlying blocks past new_size. The new size must be smaller + * than the current size. This routine can be used both for the attribute and + * data fork, and does not modify the inode size, which is left to the caller. * * The transaction passed to this routine must have made a permanent log * reservation of at least XFS_ITRUNCATE_LOG_RES. This routine may commit the @@ -1399,31 +1484,6 @@ xfs_itruncate_start( * will be "held" within the returned transaction. This routine does NOT * require any disk space to be reserved for it within the transaction. * - * The fork parameter must be either xfs_attr_fork or xfs_data_fork, and it - * indicates the fork which is to be truncated. For the attribute fork we only - * support truncation to size 0. - * - * We use the sync parameter to indicate whether or not the first transaction - * we perform might have to be synchronous. For the attr fork, it needs to be - * so if the unlink of the inode is not yet known to be permanent in the log. - * This keeps us from freeing and reusing the blocks of the attribute fork - * before the unlink of the inode becomes permanent. - * - * For the data fork, we normally have to run synchronously if we're being - * called out of the inactive path or we're being called out of the create path - * where we're truncating an existing file. Either way, the truncate needs to - * be sync so blocks don't reappear in the file with altered data in case of a - * crash. wsync filesystems can run the first case async because anything that - * shrinks the inode has to run sync so by the time we're called here from - * inactive, the inode size is permanently set to 0. - * - * Calls from the truncate path always need to be sync unless we're in a wsync - * filesystem and the file has already been unlinked. - * - * The caller is responsible for correctly setting the sync parameter. It gets - * too hard for us to guess here which path we're being called out of just - * based on inode state. - * * If we get an error, we must return with the inode locked and linked into the * current transaction. This keeps things simple for the higher level code, * because it always knows that the inode is locked and held in the transaction @@ -1431,124 +1491,34 @@ xfs_itruncate_start( * dirty on error so that transactions can be easily aborted if possible. */ int -xfs_itruncate_finish( - xfs_trans_t **tp, - xfs_inode_t *ip, - xfs_fsize_t new_size, - int fork, - int sync) +xfs_itruncate_extents( + struct xfs_trans **tpp, + struct xfs_inode *ip, + int whichfork, + xfs_fsize_t new_size) { - xfs_fsblock_t first_block; - xfs_fileoff_t first_unmap_block; - xfs_fileoff_t last_block; - xfs_filblks_t unmap_len=0; - xfs_mount_t *mp; - xfs_trans_t *ntp; - int done; - int committed; - xfs_bmap_free_t free_list; - int error; + struct xfs_mount *mp = ip->i_mount; + struct xfs_trans *tp = *tpp; + struct xfs_trans *ntp; + xfs_bmap_free_t free_list; + xfs_fsblock_t first_block; + xfs_fileoff_t first_unmap_block; + xfs_fileoff_t last_block; + xfs_filblks_t unmap_len; + int committed; + int error = 0; + int done = 0; - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL)); - ASSERT((new_size == 0) || (new_size <= ip->i_size)); - ASSERT(*tp != NULL); - ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES); - ASSERT(ip->i_transp == *tp); + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); + ASSERT(!atomic_read(&VFS_I(ip)->i_count) || + xfs_isilocked(ip, XFS_IOLOCK_EXCL)); + ASSERT(new_size <= XFS_ISIZE(ip)); + ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES); ASSERT(ip->i_itemp != NULL); ASSERT(ip->i_itemp->ili_lock_flags == 0); + ASSERT(!XFS_NOT_DQATTACHED(mp, ip)); - - ntp = *tp; - mp = (ntp)->t_mountp; - ASSERT(! XFS_NOT_DQATTACHED(mp, ip)); - - /* - * We only support truncating the entire attribute fork. - */ - if (fork == XFS_ATTR_FORK) { - new_size = 0LL; - } - first_unmap_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)new_size); - trace_xfs_itruncate_finish_start(ip, new_size); - - /* - * The first thing we do is set the size to new_size permanently - * on disk. This way we don't have to worry about anyone ever - * being able to look at the data being freed even in the face - * of a crash. What we're getting around here is the case where - * we free a block, it is allocated to another file, it is written - * to, and then we crash. If the new data gets written to the - * file but the log buffers containing the free and reallocation - * don't, then we'd end up with garbage in the blocks being freed. - * As long as we make the new_size permanent before actually - * freeing any blocks it doesn't matter if they get writtten to. - * - * The callers must signal into us whether or not the size - * setting here must be synchronous. There are a few cases - * where it doesn't have to be synchronous. Those cases - * occur if the file is unlinked and we know the unlink is - * permanent or if the blocks being truncated are guaranteed - * to be beyond the inode eof (regardless of the link count) - * and the eof value is permanent. Both of these cases occur - * only on wsync-mounted filesystems. In those cases, we're - * guaranteed that no user will ever see the data in the blocks - * that are being truncated so the truncate can run async. - * In the free beyond eof case, the file may wind up with - * more blocks allocated to it than it needs if we crash - * and that won't get fixed until the next time the file - * is re-opened and closed but that's ok as that shouldn't - * be too many blocks. - * - * However, we can't just make all wsync xactions run async - * because there's one call out of the create path that needs - * to run sync where it's truncating an existing file to size - * 0 whose size is > 0. - * - * It's probably possible to come up with a test in this - * routine that would correctly distinguish all the above - * cases from the values of the function parameters and the - * inode state but for sanity's sake, I've decided to let the - * layers above just tell us. It's simpler to correctly figure - * out in the layer above exactly under what conditions we - * can run async and I think it's easier for others read and - * follow the logic in case something has to be changed. - * cscope is your friend -- rcc. - * - * The attribute fork is much simpler. - * - * For the attribute fork we allow the caller to tell us whether - * the unlink of the inode that led to this call is yet permanent - * in the on disk log. If it is not and we will be freeing extents - * in this inode then we make the first transaction synchronous - * to make sure that the unlink is permanent by the time we free - * the blocks. - */ - if (fork == XFS_DATA_FORK) { - if (ip->i_d.di_nextents > 0) { - /* - * If we are not changing the file size then do - * not update the on-disk file size - we may be - * called from xfs_inactive_free_eofblocks(). If we - * update the on-disk file size and then the system - * crashes before the contents of the file are - * flushed to disk then the files may be full of - * holes (ie NULL files bug). - */ - if (ip->i_size != new_size) { - ip->i_d.di_size = new_size; - ip->i_size = new_size; - xfs_trans_log_inode(ntp, ip, XFS_ILOG_CORE); - } - } - } else if (sync) { - ASSERT(!(mp->m_flags & XFS_MOUNT_WSYNC)); - if (ip->i_d.di_anextents > 0) - xfs_trans_set_sync(ntp); - } - ASSERT(fork == XFS_DATA_FORK || - (fork == XFS_ATTR_FORK && - ((sync && !(mp->m_flags & XFS_MOUNT_WSYNC)) || - (sync == 0 && (mp->m_flags & XFS_MOUNT_WSYNC))))); + trace_xfs_itruncate_extents_start(ip, new_size); /* * Since it is possible for space to become allocated beyond @@ -1559,132 +1529,397 @@ xfs_itruncate_finish( * beyond the maximum file size (ie it is the same as last_block), * then there is nothing to do. */ - last_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)XFS_MAXIOFFSET(mp)); - ASSERT(first_unmap_block <= last_block); - done = 0; - if (last_block == first_unmap_block) { - done = 1; - } else { - unmap_len = last_block - first_unmap_block + 1; - } + first_unmap_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)new_size); + last_block = XFS_B_TO_FSB(mp, mp->m_super->s_maxbytes); + if (first_unmap_block == last_block) + return 0; + + ASSERT(first_unmap_block < last_block); + unmap_len = last_block - first_unmap_block + 1; while (!done) { - /* - * Free up up to XFS_ITRUNC_MAX_EXTENTS. xfs_bunmapi() - * will tell us whether it freed the entire range or - * not. If this is a synchronous mount (wsync), - * then we can tell bunmapi to keep all the - * transactions asynchronous since the unlink - * transaction that made this inode inactive has - * already hit the disk. There's no danger of - * the freed blocks being reused, there being a - * crash, and the reused blocks suddenly reappearing - * in this file with garbage in them once recovery - * runs. - */ xfs_bmap_init(&free_list, &first_block); - error = xfs_bunmapi(ntp, ip, + error = xfs_bunmapi(tp, ip, first_unmap_block, unmap_len, - xfs_bmapi_aflag(fork), + xfs_bmapi_aflag(whichfork), XFS_ITRUNC_MAX_EXTENTS, &first_block, &free_list, &done); - if (error) { - /* - * If the bunmapi call encounters an error, - * return to the caller where the transaction - * can be properly aborted. We just need to - * make sure we're not holding any resources - * that we were not when we came in. - */ - xfs_bmap_cancel(&free_list); - return error; - } + if (error) + goto out_bmap_cancel; /* * Duplicate the transaction that has the permanent * reservation and commit the old transaction. */ - error = xfs_bmap_finish(tp, &free_list, &committed); - ntp = *tp; + error = xfs_bmap_finish(&tp, &free_list, &committed); if (committed) - xfs_trans_ijoin(ntp, ip); - - if (error) { - /* - * If the bmap finish call encounters an error, return - * to the caller where the transaction can be properly - * aborted. We just need to make sure we're not - * holding any resources that we were not when we came - * in. - * - * Aborting from this point might lose some blocks in - * the file system, but oh well. - */ - xfs_bmap_cancel(&free_list); - return error; - } + xfs_trans_ijoin(tp, ip, 0); + if (error) + goto out_bmap_cancel; if (committed) { /* * Mark the inode dirty so it will be logged and * moved forward in the log as part of every commit. */ - xfs_trans_log_inode(ntp, ip, XFS_ILOG_CORE); + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); } - ntp = xfs_trans_dup(ntp); - error = xfs_trans_commit(*tp, 0); - *tp = ntp; + ntp = xfs_trans_dup(tp); + error = xfs_trans_commit(tp, 0); + tp = ntp; - xfs_trans_ijoin(ntp, ip); + xfs_trans_ijoin(tp, ip, 0); if (error) - return error; + goto out; + /* - * transaction commit worked ok so we can drop the extra ticket + * Transaction commit worked ok so we can drop the extra ticket * reference that we gained in xfs_trans_dup() */ - xfs_log_ticket_put(ntp->t_ticket); - error = xfs_trans_reserve(ntp, 0, - XFS_ITRUNCATE_LOG_RES(mp), 0, - XFS_TRANS_PERM_LOG_RES, - XFS_ITRUNCATE_LOG_COUNT); + xfs_log_ticket_put(tp->t_ticket); + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0); if (error) + goto out; + } + + /* + * Always re-log the inode so that our permanent transaction can keep + * on rolling it forward in the log. + */ + xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); + + trace_xfs_itruncate_extents_end(ip, new_size); + +out: + *tpp = tp; + return error; +out_bmap_cancel: + /* + * If the bunmapi call encounters an error, return to the caller where + * the transaction can be properly aborted. We just need to make sure + * we're not holding any resources that we were not when we came in. + */ + xfs_bmap_cancel(&free_list); + goto out; +} + +int +xfs_release( + xfs_inode_t *ip) +{ + xfs_mount_t *mp = ip->i_mount; + int error; + + if (!S_ISREG(ip->i_d.di_mode) || (ip->i_d.di_mode == 0)) + return 0; + + /* If this is a read-only mount, don't do this (would generate I/O) */ + if (mp->m_flags & XFS_MOUNT_RDONLY) + return 0; + + if (!XFS_FORCED_SHUTDOWN(mp)) { + int truncated; + + /* + * 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 + * NULL files problem which is particularly noticeable from a + * truncate down, buffered (re-)write (delalloc), followed by + * a crash. What we are effectively doing here is + * significantly reducing the time window where we'd otherwise + * be exposed to that problem. + */ + truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED); + if (truncated) { + xfs_iflags_clear(ip, XFS_IDIRTY_RELEASE); + if (VN_DIRTY(VFS_I(ip)) && ip->i_delayed_blks > 0) { + error = -filemap_flush(VFS_I(ip)->i_mapping); + if (error) + return error; + } + } + } + + if (ip->i_d.di_nlink == 0) + return 0; + + if (xfs_can_free_eofblocks(ip, false)) { + + /* + * If we can't get the iolock just skip truncating the blocks + * past EOF because we could deadlock with the mmap_sem + * otherwise. We'll get another chance to drop them once the + * last reference to the inode is dropped, so we'll never leak + * blocks permanently. + * + * Further, check if the inode is being opened, written and + * closed frequently and we have delayed allocation blocks + * outstanding (e.g. streaming writes from the NFS server), + * truncating the blocks past EOF will cause fragmentation to + * occur. + * + * In this case don't do the truncation, either, but we have to + * be careful how we detect this case. Blocks beyond EOF show + * up as i_delayed_blks even when the inode is clean, so we + * need to truncate them away first before checking for a dirty + * release. Hence on the first dirty close we will still remove + * the speculative allocation, but after that we will leave it + * in place. + */ + if (xfs_iflags_test(ip, XFS_IDIRTY_RELEASE)) + return 0; + + error = xfs_free_eofblocks(mp, ip, true); + if (error && error != EAGAIN) return error; + + /* delalloc blocks after truncation means it really is dirty */ + if (ip->i_delayed_blks) + xfs_iflags_set(ip, XFS_IDIRTY_RELEASE); + } + return 0; +} + +/* + * 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); + /* - * Only update the size in the case of the data fork, but - * always re-log the inode so that our permanent transaction - * can keep on rolling it forward in the log. + * 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. */ - if (fork == XFS_DATA_FORK) { - xfs_isize_check(mp, ip, new_size); + 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 are not changing the file size then do - * not update the on-disk file size - we may be - * called from xfs_inactive_free_eofblocks(). If we - * update the on-disk file size and then the system - * crashes before the contents of the file are - * flushed to disk then the files may be full of - * holes (ie NULL files bug). + * 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 (ip->i_size != new_size) { - ip->i_d.di_size = new_size; - ip->i_size = new_size; + 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; } - xfs_trans_log_inode(ntp, ip, XFS_ILOG_CORE); - ASSERT((new_size != 0) || - (fork == XFS_ATTR_FORK) || - (ip->i_delayed_blks == 0)); - ASSERT((new_size != 0) || - (fork == XFS_ATTR_FORK) || - (ip->i_d.di_nextents == 0)); - trace_xfs_itruncate_finish_end(ip, new_size); + + /* + * 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 + * goes to zero. If the file has been unlinked, then it must + * now be truncated. Also, we clear all of the read-ahead state + * kept for the inode here since the file is now closed. + */ +void +xfs_inactive( + xfs_inode_t *ip) +{ + struct xfs_mount *mp; + int error; + int truncate = 0; + + /* + * If the inode is already free, then there can be nothing + * to clean up here. + */ + if (ip->i_d.di_mode == 0) { + ASSERT(ip->i_df.if_real_bytes == 0); + ASSERT(ip->i_df.if_broot_bytes == 0); + return; + } + + mp = ip->i_mount; + + /* If this is a read-only mount, don't do this (would generate I/O) */ + if (mp->m_flags & XFS_MOUNT_RDONLY) + return; + + if (ip->i_d.di_nlink != 0) { + /* + * force is true because we are evicting an inode from the + * cache. Post-eof blocks must be freed, lest we end up with + * broken free space accounting. + */ + if (xfs_can_free_eofblocks(ip, true)) + xfs_free_eofblocks(mp, ip, false); + + return; + } + + if (S_ISREG(ip->i_d.di_mode) && + (ip->i_d.di_size != 0 || XFS_ISIZE(ip) != 0 || + ip->i_d.di_nextents > 0 || ip->i_delayed_blks > 0)) + truncate = 1; + + error = xfs_qm_dqattach(ip, 0); + if (error) + return; + + 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 + * now. The code calls a routine that recursively deconstructs the + * attribute fork. We need to just commit the current transaction + * because we can't use it for xfs_attr_inactive(). + */ + if (ip->i_d.di_anextents > 0) { + ASSERT(ip->i_d.di_forkoff != 0); + + error = xfs_attr_inactive(ip); + if (error) + return; + } + + if (ip->i_afp) + xfs_idestroy_fork(ip, XFS_ATTR_FORK); + + ASSERT(ip->i_d.di_anextents == 0); + + /* + * Free the inode. + */ + error = xfs_inactive_ifree(ip); + if (error) + return; + + /* + * Release the dquots held by inode, if any. + */ + xfs_qm_dqdetach(ip); +} + +/* * This is called when the inode's link count goes to 0. * We place the on-disk inode on a list in the AGI. It * will be pulled from this list when the inode is freed. @@ -1706,7 +1941,6 @@ xfs_iunlink( ASSERT(ip->i_d.di_nlink == 0); ASSERT(ip->i_d.di_mode != 0); - ASSERT(ip->i_transp == tp); mp = tp->t_mountp; @@ -1729,22 +1963,26 @@ xfs_iunlink( ASSERT(agi->agi_unlinked[bucket_index]); ASSERT(be32_to_cpu(agi->agi_unlinked[bucket_index]) != agino); - if (be32_to_cpu(agi->agi_unlinked[bucket_index]) != NULLAGINO) { + if (agi->agi_unlinked[bucket_index] != cpu_to_be32(NULLAGINO)) { /* * There is already another inode in the bucket we need * to add ourselves to. Add us at the front of the list. * Here we put the head pointer into our next pointer, * and then we fall through to point the head at us. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); + error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, + 0, 0); if (error) return error; - ASSERT(be32_to_cpu(dip->di_next_unlinked) == NULLAGINO); - /* both on-disk, don't endian flip twice */ + ASSERT(dip->di_next_unlinked == cpu_to_be32(NULLAGINO)); dip->di_next_unlinked = agi->agi_unlinked[bucket_index]; offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1806,23 +2044,22 @@ xfs_iunlink_remove( agino = XFS_INO_TO_AGINO(mp, ip->i_ino); ASSERT(agino != 0); bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS; - ASSERT(be32_to_cpu(agi->agi_unlinked[bucket_index]) != NULLAGINO); + ASSERT(agi->agi_unlinked[bucket_index] != cpu_to_be32(NULLAGINO)); ASSERT(agi->agi_unlinked[bucket_index]); if (be32_to_cpu(agi->agi_unlinked[bucket_index]) == agino) { /* - * We're at the head of the list. Get the inode's - * on-disk buffer to see if there is anyone after us - * on the list. Only modify our next pointer if it - * is not already NULLAGINO. This saves us the overhead - * of dealing with the buffer when there is no need to - * change it. + * We're at the head of the list. Get the inode's on-disk + * buffer to see if there is anyone after us on the list. + * Only modify our next pointer if it is not already NULLAGINO. + * This saves us the overhead of dealing with the buffer when + * there is no need to change it. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); + error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, + 0, 0); if (error) { - cmn_err(CE_WARN, - "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", - error, mp->m_fsname); + xfs_warn(mp, "%s: xfs_imap_to_bp returned error %d.", + __func__, error); return error; } next_agino = be32_to_cpu(dip->di_next_unlinked); @@ -1831,6 +2068,10 @@ xfs_iunlink_remove( dip->di_next_unlinked = cpu_to_be32(NULLAGINO); offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1855,36 +2096,46 @@ xfs_iunlink_remove( next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]); last_ibp = NULL; while (next_agino != agino) { - /* - * If the last inode wasn't the one pointing to - * us, then release its buffer since we're not - * going to do anything with it. - */ - if (last_ibp != NULL) { + struct xfs_imap imap; + + if (last_ibp) xfs_trans_brelse(tp, last_ibp); - } + + imap.im_blkno = 0; next_ino = XFS_AGINO_TO_INO(mp, agno, next_agino); - error = xfs_inotobp(mp, tp, next_ino, &last_dip, - &last_ibp, &last_offset, 0); + + error = xfs_imap(mp, tp, next_ino, &imap, 0); + if (error) { + xfs_warn(mp, + "%s: xfs_imap returned error %d.", + __func__, error); + return error; + } + + error = xfs_imap_to_bp(mp, tp, &imap, &last_dip, + &last_ibp, 0, 0); if (error) { - cmn_err(CE_WARN, - "xfs_iunlink_remove: xfs_inotobp() returned an error %d on %s. Returning error.", - error, mp->m_fsname); + xfs_warn(mp, + "%s: xfs_imap_to_bp returned error %d.", + __func__, error); return error; } + + last_offset = imap.im_boffset; next_agino = be32_to_cpu(last_dip->di_next_unlinked); ASSERT(next_agino != NULLAGINO); ASSERT(next_agino != 0); } + /* - * Now last_ibp points to the buffer previous to us on - * the unlinked list. Pull us from the list. + * Now last_ibp points to the buffer previous to us on the + * unlinked list. Pull us from the list. */ - error = xfs_itobp(mp, tp, ip, &dip, &ibp, XBF_LOCK); + error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, + 0, 0); if (error) { - cmn_err(CE_WARN, - "xfs_iunlink_remove: xfs_itobp() returned an error %d on %s. Returning error.", - error, mp->m_fsname); + xfs_warn(mp, "%s: xfs_imap_to_bp(2) returned error %d.", + __func__, error); return error; } next_agino = be32_to_cpu(dip->di_next_unlinked); @@ -1894,6 +2145,10 @@ xfs_iunlink_remove( dip->di_next_unlinked = cpu_to_be32(NULLAGINO); offset = ip->i_imap.im_boffset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, dip); + xfs_trans_inode_buf(tp, ibp); xfs_trans_log_buf(tp, ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1907,6 +2162,10 @@ xfs_iunlink_remove( last_dip->di_next_unlinked = cpu_to_be32(next_agino); ASSERT(next_agino != 0); offset = last_offset + offsetof(xfs_dinode_t, di_next_unlinked); + + /* need to recalc the inode CRC if appropriate */ + xfs_dinode_calc_crc(mp, last_dip); + xfs_trans_inode_buf(tp, last_ibp); xfs_trans_log_buf(tp, last_ibp, offset, (offset + sizeof(xfs_agino_t) - 1)); @@ -1916,11 +2175,11 @@ xfs_iunlink_remove( } /* - * A big issue when freeing the inode cluster is is that we _cannot_ skip any + * A big issue when freeing the inode cluster is that we _cannot_ skip any * inodes that are in memory - they all must be marked stale and attached to * the cluster buffer. */ -STATIC void +STATIC int xfs_ifree_cluster( xfs_inode_t *free_ip, xfs_trans_t *tp, @@ -1928,8 +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; @@ -1939,18 +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)); @@ -1964,7 +2216,21 @@ xfs_ifree_cluster( */ bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno, mp->m_bsize * blks_per_cluster, - XBF_LOCK); + XBF_UNMAPPED); + + if (!bp) + return ENOMEM; + + /* + * This buffer may not have been correctly initialised as we + * didn't read it from disk. That's not important because we are + * only using to mark the buffer as stale in the log, and to + * attach stale cached inodes on it. That means it will never be + * dispatched for IO. If it is, we want to know about it, and we + * want it to fail. We can acheive this by adding a write + * verifier to the buffer. + */ + bp->b_ops = &xfs_inode_buf_ops; /* * Walk the inodes already attached to the buffer and mark them @@ -1973,7 +2239,7 @@ xfs_ifree_cluster( * stale first, we will not attempt to lock them in the loop * below as the XFS_ISTALE flag will be set. */ - lip = XFS_BUF_FSPRIVATE(bp, xfs_log_item_t *); + lip = bp->b_fspriv; while (lip) { if (lip->li_type == XFS_LI_INODE) { iip = (xfs_inode_log_item_t *)lip; @@ -1998,19 +2264,35 @@ 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: - read_lock(&pag->pag_ici_lock); + rcu_read_lock(); ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, (inum + i))); - /* Inode not in memory or stale, nothing to do */ - if (!ip || xfs_iflags_test(ip, XFS_ISTALE)) { - read_unlock(&pag->pag_ici_lock); + /* Inode not in memory, nothing to do */ + if (!ip) { + rcu_read_unlock(); continue; } /* + * because this is an RCU protected lookup, we could + * find a recently freed or even reallocated inode + * during the lookup. We need to check under the + * i_flags_lock for a valid inode here. Skip it if it + * is not valid, the wrong inode or stale. + */ + spin_lock(&ip->i_flags_lock); + if (ip->i_ino != inum + i || + __xfs_iflags_test(ip, XFS_ISTALE)) { + spin_unlock(&ip->i_flags_lock); + rcu_read_unlock(); + continue; + } + spin_unlock(&ip->i_flags_lock); + + /* * Don't try to lock/unlock the current inode, but we * _cannot_ skip the other inodes that we did not find * in the list attached to the buffer and are not @@ -2019,11 +2301,11 @@ retry: */ if (ip != free_ip && !xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) { - read_unlock(&pag->pag_ici_lock); + rcu_read_unlock(); delay(1); goto retry; } - read_unlock(&pag->pag_ici_lock); + rcu_read_unlock(); xfs_iflock(ip); xfs_iflags_set(ip, XFS_ISTALE); @@ -2035,14 +2317,13 @@ retry: iip = ip->i_itemp; if (!iip || xfs_inode_clean(ip)) { ASSERT(ip != free_ip); - ip->i_update_core = 0; xfs_ifunlock(ip); xfs_iunlock(ip, XFS_ILOCK_EXCL); continue; } - iip->ili_last_fields = iip->ili_format.ilf_fields; - iip->ili_format.ilf_fields = 0; + 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); @@ -2059,6 +2340,7 @@ retry: } xfs_perag_put(pag); + return 0; } /* @@ -2080,36 +2362,29 @@ xfs_ifree( int error; int delete; xfs_ino_t first_ino; - xfs_dinode_t *dip; - xfs_buf_t *ibp; ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL)); - ASSERT(ip->i_transp == tp); ASSERT(ip->i_d.di_nlink == 0); ASSERT(ip->i_d.di_nextents == 0); ASSERT(ip->i_d.di_anextents == 0); - ASSERT((ip->i_d.di_size == 0 && ip->i_size == 0) || - ((ip->i_d.di_mode & S_IFMT) != S_IFREG)); + ASSERT(ip->i_d.di_size == 0 || !S_ISREG(ip->i_d.di_mode)); ASSERT(ip->i_d.di_nblocks == 0); /* * Pull the on-disk inode from the AGI unlinked list. */ error = xfs_iunlink_remove(tp, ip); - if (error != 0) { + if (error) return error; - } error = xfs_difree(tp, ip->i_ino, flist, &delete, &first_ino); - if (error != 0) { + if (error) return error; - } + ip->i_d.di_mode = 0; /* mark incore inode as free */ ip->i_d.di_flags = 0; ip->i_d.di_dmevmask = 0; ip->i_d.di_forkoff = 0; /* mark the attr fork not in use */ - ip->i_df.if_ext_max = - XFS_IFORK_DSIZE(ip) / (uint)sizeof(xfs_bmbt_rec_t); ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS; ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS; /* @@ -2117,489 +2392,532 @@ xfs_ifree( * by reincarnations of this inode. */ ip->i_d.di_gen++; - xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE); - error = xfs_itobp(ip->i_mount, tp, ip, &dip, &ibp, XBF_LOCK); - if (error) - return error; - - /* - * Clear the on-disk di_mode. This is to prevent xfs_bulkstat - * from picking up this inode when it is reclaimed (its incore state - * initialzed but not flushed to disk yet). The in-core di_mode is - * already cleared and a corresponding transaction logged. - * The hack here just synchronizes the in-core to on-disk - * di_mode value in advance before the actual inode sync to disk. - * This is OK because the inode is already unlinked and would never - * change its di_mode again for this inode generation. - * This is a temporary hack that would require a proper fix - * in the future. - */ - dip->di_mode = 0; - - if (delete) { - xfs_ifree_cluster(ip, tp, first_ino); - } + if (delete) + error = xfs_ifree_cluster(ip, tp, first_ino); - return 0; + return error; } /* - * Reallocate the space for if_broot based on the number of records - * being added or deleted as indicated in rec_diff. Move the records - * and pointers in if_broot to fit the new size. When shrinking this - * will eliminate holes between the records and pointers created by - * the caller. When growing this will create holes to be filled in - * by the caller. - * - * The caller must not request to add more records than would fit in - * the on-disk inode root. If the if_broot is currently NULL, then - * if we adding records one will be allocated. The caller must also - * not request that the number of records go below zero, although - * it can go to zero. - * - * ip -- the inode whose if_broot area is changing - * ext_diff -- the change in the number of records, positive or negative, - * requested for the if_broot array. + * This is called to unpin an inode. The caller must have the inode locked + * in at least shared mode so that the buffer cannot be subsequently pinned + * once someone is waiting for it to be unpinned. */ -void -xfs_iroot_realloc( - xfs_inode_t *ip, - int rec_diff, - int whichfork) +static void +xfs_iunpin( + struct xfs_inode *ip) { - struct xfs_mount *mp = ip->i_mount; - int cur_max; - xfs_ifork_t *ifp; - struct xfs_btree_block *new_broot; - int new_max; - size_t new_size; - char *np; - char *op; + ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); - /* - * Handle the degenerate case quietly. - */ - if (rec_diff == 0) { - return; - } + trace_xfs_inode_unpin_nowait(ip, _RET_IP_); - ifp = XFS_IFORK_PTR(ip, whichfork); - if (rec_diff > 0) { - /* - * If there wasn't any memory allocated before, just - * allocate it now and get out. - */ - if (ifp->if_broot_bytes == 0) { - new_size = (size_t)XFS_BMAP_BROOT_SPACE_CALC(rec_diff); - ifp->if_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS); - ifp->if_broot_bytes = (int)new_size; - return; - } + /* Give the log a push to start the unpinning I/O */ + xfs_log_force_lsn(ip->i_mount, ip->i_itemp->ili_last_lsn, 0); - /* - * If there is already an existing if_broot, then we need - * to realloc() it and shift the pointers to their new - * location. The records don't change location because - * they are kept butted up against the btree block header. - */ - cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0); - new_max = cur_max + rec_diff; - new_size = (size_t)XFS_BMAP_BROOT_SPACE_CALC(new_max); - ifp->if_broot = kmem_realloc(ifp->if_broot, new_size, - (size_t)XFS_BMAP_BROOT_SPACE_CALC(cur_max), /* old size */ - KM_SLEEP | KM_NOFS); - op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, - ifp->if_broot_bytes); - np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, - (int)new_size); - ifp->if_broot_bytes = (int)new_size; - ASSERT(ifp->if_broot_bytes <= - XFS_IFORK_SIZE(ip, whichfork) + XFS_BROOT_SIZE_ADJ); - memmove(np, op, cur_max * (uint)sizeof(xfs_dfsbno_t)); - return; - } +} - /* - * rec_diff is less than 0. In this case, we are shrinking the - * if_broot buffer. It must already exist. If we go to zero - * records, just get rid of the root and clear the status bit. - */ - ASSERT((ifp->if_broot != NULL) && (ifp->if_broot_bytes > 0)); - cur_max = xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0); - new_max = cur_max + rec_diff; - ASSERT(new_max >= 0); - if (new_max > 0) - new_size = (size_t)XFS_BMAP_BROOT_SPACE_CALC(new_max); - else - new_size = 0; - if (new_size > 0) { - new_broot = kmem_alloc(new_size, KM_SLEEP | KM_NOFS); - /* - * First copy over the btree block header. - */ - memcpy(new_broot, ifp->if_broot, XFS_BTREE_LBLOCK_LEN); - } else { - new_broot = NULL; - ifp->if_flags &= ~XFS_IFBROOT; - } +static void +__xfs_iunpin_wait( + struct xfs_inode *ip) +{ + wait_queue_head_t *wq = bit_waitqueue(&ip->i_flags, __XFS_IPINNED_BIT); + DEFINE_WAIT_BIT(wait, &ip->i_flags, __XFS_IPINNED_BIT); - /* - * Only copy the records and pointers if there are any. - */ - if (new_max > 0) { - /* - * First copy the records. - */ - op = (char *)XFS_BMBT_REC_ADDR(mp, ifp->if_broot, 1); - np = (char *)XFS_BMBT_REC_ADDR(mp, new_broot, 1); - memcpy(np, op, new_max * (uint)sizeof(xfs_bmbt_rec_t)); + xfs_iunpin(ip); - /* - * Then copy the pointers. - */ - op = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, ifp->if_broot, 1, - ifp->if_broot_bytes); - np = (char *)XFS_BMAP_BROOT_PTR_ADDR(mp, new_broot, 1, - (int)new_size); - memcpy(np, op, new_max * (uint)sizeof(xfs_dfsbno_t)); - } - kmem_free(ifp->if_broot); - ifp->if_broot = new_broot; - ifp->if_broot_bytes = (int)new_size; - ASSERT(ifp->if_broot_bytes <= - XFS_IFORK_SIZE(ip, whichfork) + XFS_BROOT_SIZE_ADJ); - return; + do { + prepare_to_wait(wq, &wait.wait, TASK_UNINTERRUPTIBLE); + if (xfs_ipincount(ip)) + io_schedule(); + } while (xfs_ipincount(ip)); + finish_wait(wq, &wait.wait); } +void +xfs_iunpin_wait( + struct xfs_inode *ip) +{ + if (xfs_ipincount(ip)) + __xfs_iunpin_wait(ip); +} /* - * This is called when the amount of space needed for if_data - * is increased or decreased. The change in size is indicated by - * the number of bytes that need to be added or deleted in the - * byte_diff parameter. + * 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. * - * If the amount of space needed has decreased below the size of the - * inline buffer, then switch to using the inline buffer. Otherwise, - * use kmem_realloc() or kmem_alloc() to adjust the size of the buffer - * to what is needed. + * 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. * - * ip -- the inode whose if_data area is changing - * byte_diff -- the change in the number of bytes, positive or negative, - * requested for the if_data array. + * 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. */ -void -xfs_idata_realloc( - xfs_inode_t *ip, - int byte_diff, - int whichfork) +int +xfs_remove( + xfs_inode_t *dp, + struct xfs_name *name, + xfs_inode_t *ip) { - xfs_ifork_t *ifp; - int new_size; - int real_size; + xfs_mount_t *mp = dp->i_mount; + xfs_trans_t *tp = NULL; + int is_dir = S_ISDIR(ip->i_d.di_mode); + int error = 0; + xfs_bmap_free_t free_list; + xfs_fsblock_t first_block; + int cancel_flags; + int committed; + int link_zero; + uint resblks; + uint log_count; + + trace_xfs_remove(dp, name); + + if (XFS_FORCED_SHUTDOWN(mp)) + return XFS_ERROR(EIO); - if (byte_diff == 0) { - return; + error = xfs_qm_dqattach(dp, 0); + if (error) + goto std_return; + + error = xfs_qm_dqattach(ip, 0); + if (error) + goto std_return; + + if (is_dir) { + tp = xfs_trans_alloc(mp, XFS_TRANS_RMDIR); + log_count = XFS_DEFAULT_LOG_COUNT; + } else { + tp = xfs_trans_alloc(mp, XFS_TRANS_REMOVE); + log_count = XFS_REMOVE_LOG_COUNT; + } + cancel_flags = XFS_TRANS_RELEASE_LOG_RES; + + /* + * We try to get the real space reservation first, + * allowing for directory btree deletion(s) implying + * possible bmap insert(s). If we can't get the space + * reservation then we use 0 instead, and avoid the bmap + * btree insert(s) in the directory code by, if the bmap + * insert tries to happen, instead trimming the LAST + * block from the directory. + */ + resblks = XFS_REMOVE_SPACE_RES(mp); + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, resblks, 0); + if (error == ENOSPC) { + resblks = 0; + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_remove, 0, 0); } + if (error) { + ASSERT(error != ENOSPC); + cancel_flags = 0; + goto out_trans_cancel; + } + + xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL); - ifp = XFS_IFORK_PTR(ip, whichfork); - new_size = (int)ifp->if_bytes + byte_diff; - ASSERT(new_size >= 0); + xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL); - if (new_size == 0) { - if (ifp->if_u1.if_data != ifp->if_u2.if_inline_data) { - kmem_free(ifp->if_u1.if_data); + /* + * 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) { + error = XFS_ERROR(ENOTEMPTY); + goto out_trans_cancel; } - ifp->if_u1.if_data = NULL; - real_size = 0; - } else if (new_size <= sizeof(ifp->if_u2.if_inline_data)) { - /* - * If the valid extents/data can fit in if_inline_ext/data, - * copy them from the malloc'd vector and free it. - */ - if (ifp->if_u1.if_data == NULL) { - ifp->if_u1.if_data = ifp->if_u2.if_inline_data; - } else if (ifp->if_u1.if_data != ifp->if_u2.if_inline_data) { - ASSERT(ifp->if_real_bytes != 0); - memcpy(ifp->if_u2.if_inline_data, ifp->if_u1.if_data, - new_size); - kmem_free(ifp->if_u1.if_data); - ifp->if_u1.if_data = ifp->if_u2.if_inline_data; + if (!xfs_dir_isempty(ip)) { + error = XFS_ERROR(ENOTEMPTY); + goto out_trans_cancel; } - real_size = 0; + + /* Drop the link from ip's "..". */ + error = xfs_droplink(tp, dp); + if (error) + goto out_trans_cancel; + + /* Drop the "." link from ip to self. */ + error = xfs_droplink(tp, ip); + if (error) + goto out_trans_cancel; } else { /* - * Stuck with malloc/realloc. - * For inline data, the underlying buffer must be - * a multiple of 4 bytes in size so that it can be - * logged and stay on word boundaries. We enforce - * that here. + * When removing a non-directory we need to log the parent + * inode here. For a directory this is done implicitly + * by the xfs_droplink call for the ".." entry. */ - real_size = roundup(new_size, 4); - if (ifp->if_u1.if_data == NULL) { - ASSERT(ifp->if_real_bytes == 0); - ifp->if_u1.if_data = kmem_alloc(real_size, - KM_SLEEP | KM_NOFS); - } else if (ifp->if_u1.if_data != ifp->if_u2.if_inline_data) { - /* - * Only do the realloc if the underlying size - * is really changing. - */ - if (ifp->if_real_bytes != real_size) { - ifp->if_u1.if_data = - kmem_realloc(ifp->if_u1.if_data, - real_size, - ifp->if_real_bytes, - KM_SLEEP | KM_NOFS); - } - } else { - ASSERT(ifp->if_real_bytes == 0); - ifp->if_u1.if_data = kmem_alloc(real_size, - KM_SLEEP | KM_NOFS); - memcpy(ifp->if_u1.if_data, ifp->if_u2.if_inline_data, - ifp->if_bytes); - } + xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE); } - ifp->if_real_bytes = real_size; - ifp->if_bytes = new_size; - ASSERT(ifp->if_bytes <= XFS_IFORK_SIZE(ip, whichfork)); + xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + + /* Drop the link from dp to ip. */ + error = xfs_droplink(tp, ip); + if (error) + goto out_trans_cancel; + + /* 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 + * the user. + */ + if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) + xfs_trans_set_sync(tp); + + error = xfs_bmap_finish(&tp, &free_list, &committed); + if (error) + goto out_bmap_cancel; + + error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + if (error) + goto std_return; + + if (is_dir && xfs_inode_is_filestream(ip)) + xfs_filestream_deassociate(ip); + + return 0; + + out_bmap_cancel: + xfs_bmap_cancel(&free_list); + out_trans_cancel: + xfs_trans_cancel(tp, cancel_flags); + std_return: + return error; } -void -xfs_idestroy_fork( - xfs_inode_t *ip, - int whichfork) +/* + * Enter all inodes for a rename transaction into a sorted array. + */ +STATIC void +xfs_sort_for_rename( + xfs_inode_t *dp1, /* in: old (source) directory inode */ + xfs_inode_t *dp2, /* in: new (target) directory inode */ + xfs_inode_t *ip1, /* in: inode of old entry */ + xfs_inode_t *ip2, /* in: inode of new entry, if it + already exists, NULL otherwise. */ + xfs_inode_t **i_tab,/* out: array of inode returned, sorted */ + int *num_inodes) /* out: number of inodes in array */ { - xfs_ifork_t *ifp; + xfs_inode_t *temp; + int i, j; - ifp = XFS_IFORK_PTR(ip, whichfork); - if (ifp->if_broot != NULL) { - kmem_free(ifp->if_broot); - ifp->if_broot = NULL; + /* + * i_tab contains a list of pointers to inodes. We initialize + * the table here & we'll sort it. We will then use it to + * order the acquisition of the inode locks. + * + * Note that the table may contain duplicates. e.g., dp1 == dp2. + */ + i_tab[0] = dp1; + i_tab[1] = dp2; + i_tab[2] = ip1; + if (ip2) { + *num_inodes = 4; + i_tab[3] = ip2; + } else { + *num_inodes = 3; + i_tab[3] = NULL; } /* - * If the format is local, then we can't have an extents - * array so just look for an inline data array. If we're - * not local then we may or may not have an extents list, - * so check and free it up if we do. + * Sort the elements via bubble sort. (Remember, there are at + * most 4 elements to sort, so this is adequate.) */ - if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) { - if ((ifp->if_u1.if_data != ifp->if_u2.if_inline_data) && - (ifp->if_u1.if_data != NULL)) { - ASSERT(ifp->if_real_bytes != 0); - kmem_free(ifp->if_u1.if_data); - ifp->if_u1.if_data = NULL; - ifp->if_real_bytes = 0; + for (i = 0; i < *num_inodes; i++) { + for (j = 1; j < *num_inodes; j++) { + if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) { + temp = i_tab[j]; + i_tab[j] = i_tab[j-1]; + i_tab[j-1] = temp; + } } - } else if ((ifp->if_flags & XFS_IFEXTENTS) && - ((ifp->if_flags & XFS_IFEXTIREC) || - ((ifp->if_u1.if_extents != NULL) && - (ifp->if_u1.if_extents != ifp->if_u2.if_inline_ext)))) { - ASSERT(ifp->if_real_bytes != 0); - xfs_iext_destroy(ifp); - } - ASSERT(ifp->if_u1.if_extents == NULL || - ifp->if_u1.if_extents == ifp->if_u2.if_inline_ext); - ASSERT(ifp->if_real_bytes == 0); - if (whichfork == XFS_ATTR_FORK) { - kmem_zone_free(xfs_ifork_zone, ip->i_afp); - ip->i_afp = NULL; } } /* - * This is called to unpin an inode. The caller must have the inode locked - * in at least shared mode so that the buffer cannot be subsequently pinned - * once someone is waiting for it to be unpinned. + * xfs_rename */ -static void -xfs_iunpin_nowait( - struct xfs_inode *ip) +int +xfs_rename( + xfs_inode_t *src_dp, + struct xfs_name *src_name, + xfs_inode_t *src_ip, + xfs_inode_t *target_dp, + struct xfs_name *target_name, + xfs_inode_t *target_ip) { - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); + xfs_trans_t *tp = NULL; + xfs_mount_t *mp = src_dp->i_mount; + int new_parent; /* moving to a new dir */ + int src_is_directory; /* src_name is a directory */ + int error; + xfs_bmap_free_t free_list; + xfs_fsblock_t first_block; + int cancel_flags; + int committed; + xfs_inode_t *inodes[4]; + int spaceres; + int num_inodes; - trace_xfs_inode_unpin_nowait(ip, _RET_IP_); + trace_xfs_rename(src_dp, target_dp, src_name, target_name); - /* Give the log a push to start the unpinning I/O */ - xfs_log_force_lsn(ip->i_mount, ip->i_itemp->ili_last_lsn, 0); + new_parent = (src_dp != target_dp); + src_is_directory = S_ISDIR(src_ip->i_d.di_mode); -} + xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, + inodes, &num_inodes); -void -xfs_iunpin_wait( - struct xfs_inode *ip) -{ - if (xfs_ipincount(ip)) { - xfs_iunpin_nowait(ip); - wait_event(ip->i_ipin_wait, (xfs_ipincount(ip) == 0)); + xfs_bmap_init(&free_list, &first_block); + tp = xfs_trans_alloc(mp, XFS_TRANS_RENAME); + cancel_flags = XFS_TRANS_RELEASE_LOG_RES; + spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len); + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, spaceres, 0); + if (error == ENOSPC) { + spaceres = 0; + error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, 0, 0); + } + if (error) { + xfs_trans_cancel(tp, 0); + goto std_return; } -} -/* - * xfs_iextents_copy() - * - * This is called to copy the REAL extents (as opposed to the delayed - * allocation extents) from the inode into the given buffer. It - * returns the number of bytes copied into the buffer. - * - * If there are no delayed allocation extents, then we can just - * memcpy() the extents into the buffer. Otherwise, we need to - * examine each extent in turn and skip those which are delayed. - */ -int -xfs_iextents_copy( - xfs_inode_t *ip, - xfs_bmbt_rec_t *dp, - int whichfork) -{ - int copied; - int i; - xfs_ifork_t *ifp; - int nrecs; - xfs_fsblock_t start_block; + /* + * Attach the dquots to the inodes + */ + error = xfs_qm_vop_rename_dqattach(inodes); + if (error) { + xfs_trans_cancel(tp, cancel_flags); + goto std_return; + } - ifp = XFS_IFORK_PTR(ip, whichfork); - ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); - ASSERT(ifp->if_bytes > 0); + /* + * Lock all the participating inodes. Depending upon whether + * the target_name exists in the target directory, and + * whether the target directory is the same as the source + * directory, we can lock from 2 to 4 inodes. + */ + xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL); - nrecs = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - XFS_BMAP_TRACE_EXLIST(ip, nrecs, whichfork); - ASSERT(nrecs > 0); + /* + * Join all the inodes to the transaction. From this point on, + * we can rely on either trans_commit or trans_cancel to unlock + * them. + */ + xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL); + if (new_parent) + xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL); + xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL); + if (target_ip) + xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL); /* - * There are some delayed allocation extents in the - * inode, so copy the extents one at a time and skip - * the delayed ones. There must be at least one - * non-delayed extent. + * If we are using project inheritance, we only allow renames + * into our tree when the project IDs are the same; else the + * tree quota mechanism would be circumvented. */ - copied = 0; - for (i = 0; i < nrecs; i++) { - xfs_bmbt_rec_host_t *ep = xfs_iext_get_ext(ifp, i); - start_block = xfs_bmbt_get_startblock(ep); - if (isnullstartblock(start_block)) { + if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) && + (xfs_get_projid(target_dp) != xfs_get_projid(src_ip)))) { + error = XFS_ERROR(EXDEV); + goto error_return; + } + + /* + * Set up the target. + */ + if (target_ip == NULL) { + /* + * If there's no space reservation, check the entry will + * fit before actually inserting it. + */ + error = xfs_dir_canenter(tp, target_dp, target_name, spaceres); + if (error) + goto error_return; + /* + * If target does not exist and the rename crosses + * directories, adjust the target directory link count + * to account for the ".." reference from the new entry. + */ + error = xfs_dir_createname(tp, target_dp, target_name, + src_ip->i_ino, &first_block, + &free_list, spaceres); + if (error == ENOSPC) + goto error_return; + if (error) + goto abort_return; + + xfs_trans_ichgtime(tp, target_dp, + XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + + if (new_parent && src_is_directory) { + error = xfs_bumplink(tp, target_dp); + if (error) + goto abort_return; + } + } else { /* target_ip != NULL */ + /* + * If target exists and it's a directory, check that both + * target and source are directories and that target can be + * destroyed, or that neither is a directory. + */ + if (S_ISDIR(target_ip->i_d.di_mode)) { /* - * It's a delayed allocation extent, so skip it. + * Make sure target dir is empty. */ - continue; + if (!(xfs_dir_isempty(target_ip)) || + (target_ip->i_d.di_nlink > 2)) { + error = XFS_ERROR(EEXIST); + goto error_return; + } } - /* Translate to on disk format */ - put_unaligned(cpu_to_be64(ep->l0), &dp->l0); - put_unaligned(cpu_to_be64(ep->l1), &dp->l1); - dp++; - copied++; - } - ASSERT(copied != 0); - xfs_validate_extents(ifp, copied, XFS_EXTFMT_INODE(ip)); + /* + * Link the source inode under the target name. + * If the source inode is a directory and we are moving + * it across directories, its ".." entry will be + * inconsistent until we replace that down below. + * + * In case there is already an entry with the same + * name at the destination directory, remove it first. + */ + error = xfs_dir_replace(tp, target_dp, target_name, + src_ip->i_ino, + &first_block, &free_list, spaceres); + if (error) + goto abort_return; - return (copied * (uint)sizeof(xfs_bmbt_rec_t)); -} + xfs_trans_ichgtime(tp, target_dp, + XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + + /* + * Decrement the link count on the target since the target + * dir no longer points to it. + */ + error = xfs_droplink(tp, target_ip); + if (error) + goto abort_return; + + if (src_is_directory) { + /* + * Drop the link from the old "." entry. + */ + error = xfs_droplink(tp, target_ip); + if (error) + goto abort_return; + } + } /* target_ip != NULL */ -/* - * Each of the following cases stores data into the same region - * of the on-disk inode, so only one of them can be valid at - * any given time. While it is possible to have conflicting formats - * and log flags, e.g. having XFS_ILOG_?DATA set when the fork is - * in EXTENTS format, this can only happen when the fork has - * changed formats after being modified but before being flushed. - * In these cases, the format always takes precedence, because the - * format indicates the current state of the fork. - */ -/*ARGSUSED*/ -STATIC void -xfs_iflush_fork( - xfs_inode_t *ip, - xfs_dinode_t *dip, - xfs_inode_log_item_t *iip, - int whichfork, - xfs_buf_t *bp) -{ - char *cp; - xfs_ifork_t *ifp; - xfs_mount_t *mp; -#ifdef XFS_TRANS_DEBUG - int first; -#endif - static const short brootflag[2] = - { XFS_ILOG_DBROOT, XFS_ILOG_ABROOT }; - static const short dataflag[2] = - { XFS_ILOG_DDATA, XFS_ILOG_ADATA }; - static const short extflag[2] = - { XFS_ILOG_DEXT, XFS_ILOG_AEXT }; - - if (!iip) - return; - ifp = XFS_IFORK_PTR(ip, whichfork); /* - * This can happen if we gave up in iformat in an error path, - * for the attribute fork. + * Remove the source. */ - if (!ifp) { - ASSERT(whichfork == XFS_ATTR_FORK); - return; + if (new_parent && src_is_directory) { + /* + * Rewrite the ".." entry to point to the new + * directory. + */ + error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot, + target_dp->i_ino, + &first_block, &free_list, spaceres); + ASSERT(error != EEXIST); + if (error) + goto abort_return; } - cp = XFS_DFORK_PTR(dip, whichfork); - mp = ip->i_mount; - switch (XFS_IFORK_FORMAT(ip, whichfork)) { - case XFS_DINODE_FMT_LOCAL: - if ((iip->ili_format.ilf_fields & dataflag[whichfork]) && - (ifp->if_bytes > 0)) { - ASSERT(ifp->if_u1.if_data != NULL); - ASSERT(ifp->if_bytes <= XFS_IFORK_SIZE(ip, whichfork)); - memcpy(cp, ifp->if_u1.if_data, ifp->if_bytes); - } - break; - case XFS_DINODE_FMT_EXTENTS: - ASSERT((ifp->if_flags & XFS_IFEXTENTS) || - !(iip->ili_format.ilf_fields & extflag[whichfork])); - ASSERT((xfs_iext_get_ext(ifp, 0) != NULL) || - (ifp->if_bytes == 0)); - ASSERT((xfs_iext_get_ext(ifp, 0) == NULL) || - (ifp->if_bytes > 0)); - if ((iip->ili_format.ilf_fields & extflag[whichfork]) && - (ifp->if_bytes > 0)) { - ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) > 0); - (void)xfs_iextents_copy(ip, (xfs_bmbt_rec_t *)cp, - whichfork); - } - break; + /* + * We always want to hit the ctime on the source inode. + * + * This isn't strictly required by the standards since the source + * inode isn't really being changed, but old unix file systems did + * it and some incremental backup programs won't work without it. + */ + xfs_trans_ichgtime(tp, src_ip, XFS_ICHGTIME_CHG); + xfs_trans_log_inode(tp, src_ip, XFS_ILOG_CORE); - case XFS_DINODE_FMT_BTREE: - if ((iip->ili_format.ilf_fields & brootflag[whichfork]) && - (ifp->if_broot_bytes > 0)) { - ASSERT(ifp->if_broot != NULL); - ASSERT(ifp->if_broot_bytes <= - (XFS_IFORK_SIZE(ip, whichfork) + - XFS_BROOT_SIZE_ADJ)); - xfs_bmbt_to_bmdr(mp, ifp->if_broot, ifp->if_broot_bytes, - (xfs_bmdr_block_t *)cp, - XFS_DFORK_SIZE(dip, mp, whichfork)); - } - break; + /* + * Adjust the link count on src_dp. This is necessary when + * renaming a directory, either within one parent when + * the target existed, or across two parent directories. + */ + if (src_is_directory && (new_parent || target_ip != NULL)) { - case XFS_DINODE_FMT_DEV: - if (iip->ili_format.ilf_fields & XFS_ILOG_DEV) { - ASSERT(whichfork == XFS_DATA_FORK); - xfs_dinode_put_rdev(dip, ip->i_df.if_u2.if_rdev); - } - break; + /* + * Decrement link count on src_directory since the + * entry that's moved no longer points to it. + */ + error = xfs_droplink(tp, src_dp); + if (error) + goto abort_return; + } - case XFS_DINODE_FMT_UUID: - if (iip->ili_format.ilf_fields & XFS_ILOG_UUID) { - ASSERT(whichfork == XFS_DATA_FORK); - memcpy(XFS_DFORK_DPTR(dip), - &ip->i_df.if_u2.if_uuid, - sizeof(uuid_t)); - } - break; + error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino, + &first_block, &free_list, spaceres); + if (error) + goto abort_return; - default: - ASSERT(0); - break; + xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); + xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE); + if (new_parent) + xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE); + + /* + * If this is a synchronous mount, make sure that the + * rename transaction goes to disk before returning to + * the user. + */ + if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC)) { + xfs_trans_set_sync(tp); + } + + error = xfs_bmap_finish(&tp, &free_list, &committed); + if (error) { + xfs_bmap_cancel(&free_list); + xfs_trans_cancel(tp, (XFS_TRANS_RELEASE_LOG_RES | + XFS_TRANS_ABORT)); + goto std_return; } + + /* + * trans_commit will unlock src_ip, target_ip & decrement + * the vnode references. + */ + return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES); + + abort_return: + cancel_flags |= XFS_TRANS_ABORT; + error_return: + xfs_bmap_cancel(&free_list); + xfs_trans_cancel(tp, cancel_flags); + std_return: + return error; } STATIC int @@ -2621,15 +2939,15 @@ 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; - read_lock(&pag->pag_ici_lock); + rcu_read_lock(); /* really need a gang lookup range call here */ nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)ilist, first_index, inodes_per_cluster); @@ -2640,9 +2958,21 @@ xfs_iflush_cluster( iq = ilist[i]; if (iq == ip) continue; - /* if the inode lies outside this cluster, we're done. */ - if ((XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) - break; + + /* + * because this is an RCU protected lookup, we could find a + * recently freed or even reallocated inode during the lookup. + * We need to check under the i_flags_lock for a valid inode + * here. Skip it if it is not valid or the wrong inode. + */ + spin_lock(&ip->i_flags_lock); + if (!ip->i_ino || + (XFS_INO_TO_AGINO(mp, iq->i_ino) & mask) != first_index) { + spin_unlock(&ip->i_flags_lock); + continue; + } + spin_unlock(&ip->i_flags_lock); + /* * Do an un-protected check to see if the inode is dirty and * is a candidate for flushing. These checks will be repeated @@ -2692,7 +3022,7 @@ xfs_iflush_cluster( } out_free: - read_unlock(&pag->pag_ici_lock); + rcu_read_unlock(); kmem_free(ilist); out_put: xfs_perag_put(pag); @@ -2704,13 +3034,13 @@ cluster_corrupt_out: * Corruption detected in the clustering loop. Invalidate the * inode buffer and shut down the filesystem. */ - read_unlock(&pag->pag_ici_lock); + rcu_read_unlock(); /* - * Clean up the buffer. If it was B_DELWRI, just release it -- + * Clean up the buffer. If it was delwri, just release it -- * brelse can handle it with no problems. If not, shut down the * filesystem before releasing the buffer. */ - bufwasdelwri = XFS_BUF_ISDELAYWRITE(bp); + bufwasdelwri = (bp->b_flags & _XBF_DELWRI_Q); if (bufwasdelwri) xfs_buf_relse(bp); @@ -2722,13 +3052,13 @@ cluster_corrupt_out: * mark the buffer as an error and call them. Otherwise * mark it as stale and brelse. */ - if (XFS_BUF_IODONE_FUNC(bp)) { + if (bp->b_iodone) { XFS_BUF_UNDONE(bp); - XFS_BUF_STALE(bp); - XFS_BUF_ERROR(bp,EIO); + xfs_buf_stale(bp); + xfs_buf_ioerror(bp, EIO); xfs_buf_ioend(bp, 0); } else { - XFS_BUF_STALE(bp); + xfs_buf_stale(bp); xfs_buf_relse(bp); } } @@ -2736,64 +3066,46 @@ cluster_corrupt_out: /* * Unlocks the flush lock */ - xfs_iflush_abort(iq); + xfs_iflush_abort(iq, false); kmem_free(ilist); xfs_perag_put(pag); return XFS_ERROR(EFSCORRUPTED); } /* - * xfs_iflush() will write a modified inode's changes out to the - * inode's on disk home. The caller must have the inode lock held - * in at least shared mode and the inode flush completion must be - * active as well. The inode lock will still be held upon return from - * the call and the caller is free to unlock it. - * The inode flush will be completed when the inode reaches the disk. - * The flags indicate how the inode's buffer should be written out. + * Flush dirty inode metadata into the backing buffer. + * + * The caller must have the inode lock and the inode flush lock held. The + * inode lock will still be held upon return to the caller, and the inode + * flush lock will be released after the inode has reached the disk. + * + * The caller must write out the buffer returned in *bpp and release it. */ int xfs_iflush( - xfs_inode_t *ip, - uint flags) + struct xfs_inode *ip, + struct xfs_buf **bpp) { - xfs_inode_log_item_t *iip; - xfs_buf_t *bp; - xfs_dinode_t *dip; - xfs_mount_t *mp; + struct xfs_mount *mp = ip->i_mount; + struct xfs_buf *bp; + struct xfs_dinode *dip; int error; XFS_STATS_INC(xs_iflush_count); ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)); - ASSERT(!completion_done(&ip->i_flush)); + ASSERT(xfs_isiflocked(ip)); ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE || - ip->i_d.di_nextents > ip->i_df.if_ext_max); + ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK)); - iip = ip->i_itemp; - mp = ip->i_mount; + *bpp = NULL; - /* - * We can't flush the inode until it is unpinned, so wait for it if we - * are allowed to block. We know noone new can pin it, because we are - * holding the inode lock shared and you need to hold it exclusively to - * pin the inode. - * - * If we are not allowed to block, force the log out asynchronously so - * that when we come back the inode will be unpinned. If other inodes - * in the same cluster are dirty, they will probably write the inode - * out for us if they occur after the log force completes. - */ - if (!(flags & SYNC_WAIT) && xfs_ipincount(ip)) { - xfs_iunpin_nowait(ip); - xfs_ifunlock(ip); - return EAGAIN; - } xfs_iunpin_wait(ip); /* * For stale inodes we cannot rely on the backing buffer remaining * stale in cache for the remaining life of the stale inode and so - * xfs_itobp() below may give us a buffer that no longer contains + * xfs_imap_to_bp() below may give us a buffer that no longer contains * inodes below. We have to check this after ensuring the inode is * unpinned so that it is safe to reclaim the stale inode after the * flush call. @@ -2806,21 +3118,21 @@ xfs_iflush( /* * This may have been unpinned because the filesystem is shutting * down forcibly. If that's the case we must not write this inode - * to disk, because the log record didn't make it to disk! + * to disk, because the log record didn't make it to disk. + * + * We also have to remove the log item from the AIL in this case, + * as we wait for an empty AIL as part of the unmount process. */ if (XFS_FORCED_SHUTDOWN(mp)) { - ip->i_update_core = 0; - if (iip) - iip->ili_format.ilf_fields = 0; - xfs_ifunlock(ip); - return XFS_ERROR(EIO); + error = XFS_ERROR(EIO); + goto abort_out; } /* * Get the buffer containing the on-disk inode. */ - error = xfs_itobp(mp, NULL, ip, &dip, &bp, - (flags & SYNC_WAIT) ? XBF_LOCK : XBF_TRYLOCK); + error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &bp, XBF_TRYLOCK, + 0); if (error || !bp) { xfs_ifunlock(ip); return error; @@ -2837,7 +3149,7 @@ xfs_iflush( * If the buffer is pinned then push on the log now so we won't * get stuck waiting in the write for too long. */ - if (XFS_BUF_ISPINNED(bp)) + if (xfs_buf_ispinned(bp)) xfs_log_force(mp, 0); /* @@ -2848,127 +3160,107 @@ xfs_iflush( if (error) goto cluster_corrupt_out; - if (flags & SYNC_WAIT) - error = xfs_bwrite(mp, bp); - else - xfs_bdwrite(mp, bp); - return error; + *bpp = bp; + return 0; corrupt_out: xfs_buf_relse(bp); xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE); cluster_corrupt_out: + error = XFS_ERROR(EFSCORRUPTED); +abort_out: /* * Unlocks the flush lock */ - xfs_iflush_abort(ip); - return XFS_ERROR(EFSCORRUPTED); + xfs_iflush_abort(ip, false); + return error; } - 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; -#ifdef XFS_TRANS_DEBUG - int first; -#endif + 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(!completion_done(&ip->i_flush)); + ASSERT(xfs_isiflocked(ip)); ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE || - ip->i_d.di_nextents > ip->i_df.if_ext_max); - - iip = ip->i_itemp; - mp = ip->i_mount; + 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); - /* - * Clear i_update_core before copying out the data. - * This is for coordination with our timestamp updates - * that don't hold the inode lock. They will always - * update the timestamps BEFORE setting i_update_core, - * so if we clear i_update_core after they set it we - * are guaranteed to see their updates to the timestamps. - * I believe that this depends on strongly ordered memory - * semantics, but we have that. We use the SYNCHRONIZE - * macro to make sure that the compiler does not reorder - * the i_update_core access below the data copy below. - */ - ip->i_update_core = 0; - SYNCHRONIZE(); - - /* - * Make sure to get the latest timestamps from the Linux inode. - */ - xfs_synchronize_times(ip); - - if (XFS_TEST_ERROR(be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC, + if (XFS_TEST_ERROR(dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC), mp, XFS_ERRTAG_IFLUSH_1, XFS_RANDOM_IFLUSH_1)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: Bad inode %Lu magic number 0x%x, ptr 0x%p", - ip->i_ino, be16_to_cpu(dip->di_magic), dip); + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: Bad inode %Lu magic number 0x%x, ptr 0x%p", + __func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip); goto corrupt_out; } if (XFS_TEST_ERROR(ip->i_d.di_magic != XFS_DINODE_MAGIC, mp, XFS_ERRTAG_IFLUSH_2, XFS_RANDOM_IFLUSH_2)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: Bad inode %Lu, ptr 0x%p, magic number 0x%x", - ip->i_ino, ip, ip->i_d.di_magic); + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: Bad inode %Lu, ptr 0x%p, magic number 0x%x", + __func__, ip->i_ino, ip, ip->i_d.di_magic); goto corrupt_out; } - if ((ip->i_d.di_mode & S_IFMT) == S_IFREG) { + if (S_ISREG(ip->i_d.di_mode)) { if (XFS_TEST_ERROR( (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) && (ip->i_d.di_format != XFS_DINODE_FMT_BTREE), mp, XFS_ERRTAG_IFLUSH_3, XFS_RANDOM_IFLUSH_3)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: Bad regular inode %Lu, ptr 0x%p", - ip->i_ino, ip); + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: Bad regular inode %Lu, ptr 0x%p", + __func__, ip->i_ino, ip); goto corrupt_out; } - } else if ((ip->i_d.di_mode & S_IFMT) == S_IFDIR) { + } else if (S_ISDIR(ip->i_d.di_mode)) { if (XFS_TEST_ERROR( (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) && (ip->i_d.di_format != XFS_DINODE_FMT_BTREE) && (ip->i_d.di_format != XFS_DINODE_FMT_LOCAL), mp, XFS_ERRTAG_IFLUSH_4, XFS_RANDOM_IFLUSH_4)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: Bad directory inode %Lu, ptr 0x%p", - ip->i_ino, ip); + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: Bad directory inode %Lu, ptr 0x%p", + __func__, ip->i_ino, ip); goto corrupt_out; } } if (XFS_TEST_ERROR(ip->i_d.di_nextents + ip->i_d.di_anextents > ip->i_d.di_nblocks, mp, XFS_ERRTAG_IFLUSH_5, XFS_RANDOM_IFLUSH_5)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: detected corrupt incore inode %Lu, total extents = %d, nblocks = %Ld, ptr 0x%p", - ip->i_ino, + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: detected corrupt incore inode %Lu, " + "total extents = %d, nblocks = %Ld, ptr 0x%p", + __func__, ip->i_ino, ip->i_d.di_nextents + ip->i_d.di_anextents, - ip->i_d.di_nblocks, - ip); + ip->i_d.di_nblocks, ip); goto corrupt_out; } if (XFS_TEST_ERROR(ip->i_d.di_forkoff > mp->m_sb.sb_inodesize, mp, XFS_ERRTAG_IFLUSH_6, XFS_RANDOM_IFLUSH_6)) { - xfs_cmn_err(XFS_PTAG_IFLUSH, CE_ALERT, mp, - "xfs_iflush: bad inode %Lu, forkoff 0x%x, ptr 0x%p", - ip->i_ino, ip->i_d.di_forkoff, ip); + xfs_alert_tag(mp, XFS_PTAG_IFLUSH, + "%s: bad inode %Lu, forkoff 0x%x, ptr 0x%p", + __func__, ip->i_ino, ip->i_d.di_forkoff, ip); goto corrupt_out; } + /* - * bump the flush iteration count, used to detect flushes which - * postdate a log record during recovery. + * 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 + * hence this can't happen but we need to still do it to ensure + * backwards compatibility with old kernels that predate logging all + * inode changes. */ - - ip->i_d.di_flushiter++; + if (ip->i_d.di_version < 3) + ip->i_d.di_flushiter++; /* * Copy the dirty parts of the inode into the on-disk @@ -2982,1137 +3274,62 @@ 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); /* - * We've recorded everything logged in the inode, so we'd - * like to clear the ilf_fields bits so we don't log and - * flush things unnecessarily. However, we can't stop - * logging all this information until the data we've copied - * into the disk buffer is written to disk. If we did we might - * overwrite the copy of the inode in the log with all the - * data after re-logging only part of it, and in the face of - * a crash we wouldn't have all the data we need to recover. + * We've recorded everything logged in the inode, so we'd like to clear + * the ili_fields bits so we don't log and flush things unnecessarily. + * However, we can't stop logging all this information until the data + * we've copied into the disk buffer is written to disk. If we did we + * might overwrite the copy of the inode in the log with all the data + * after re-logging only part of it, and in the face of a crash we + * wouldn't have all the data we need to recover. * - * What we do is move the bits to the ili_last_fields field. - * When logging the inode, these bits are moved back to the - * ilf_fields field. In the xfs_iflush_done() routine we - * clear ili_last_fields, since we know that the information - * those bits represent is permanently on disk. As long as - * the flush completes before the inode is logged again, then - * both ilf_fields and ili_last_fields will be cleared. + * What we do is move the bits to the ili_last_fields field. When + * logging the inode, these bits are moved back to the ili_fields field. + * In the xfs_iflush_done() routine we clear ili_last_fields, since we + * know that the information those bits represent is permanently on + * disk. As long as the flush completes before the inode is logged + * again, then both ili_fields and ili_last_fields will be cleared. * - * We can play with the ilf_fields bits here, because the inode - * lock must be held exclusively in order to set bits there - * and the flush lock protects the ili_last_fields bits. - * Set ili_logged so the flush done - * routine can tell whether or not to look in the AIL. - * Also, store the current LSN of the inode so that we can tell - * whether the item has moved in the AIL from xfs_iflush_done(). - * In order to read the lsn we need the AIL lock, because - * it is a 64 bit value that cannot be read atomically. + * We can play with the ili_fields bits here, because the inode lock + * must be held exclusively in order to set bits there and the flush + * lock protects the ili_last_fields bits. Set ili_logged so the flush + * done routine can tell whether or not to look in the AIL. Also, store + * the current LSN of the inode so that we can tell whether the item has + * moved in the AIL from xfs_iflush_done(). In order to read the lsn we + * need the AIL lock, because it is a 64 bit value that cannot be read + * atomically. */ - if (iip != NULL && iip->ili_format.ilf_fields != 0) { - iip->ili_last_fields = iip->ili_format.ilf_fields; - iip->ili_format.ilf_fields = 0; - iip->ili_logged = 1; - - 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); - - ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL); - ASSERT(XFS_BUF_IODONE_FUNC(bp) != NULL); - } else { - /* - * We're flushing an inode which is not in the AIL and has - * not been logged but has i_update_core set. For this - * case we can use a B_DELWRI flush and 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); - } + iip->ili_last_fields = iip->ili_fields; + iip->ili_fields = 0; + iip->ili_logged = 1; - return 0; - -corrupt_out: - return XFS_ERROR(EFSCORRUPTED); -} - -/* - * Return a pointer to the extent record at file index idx. - */ -xfs_bmbt_rec_host_t * -xfs_iext_get_ext( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t idx) /* index of target extent */ -{ - ASSERT(idx >= 0); - if ((ifp->if_flags & XFS_IFEXTIREC) && (idx == 0)) { - return ifp->if_u1.if_ext_irec->er_extbuf; - } else if (ifp->if_flags & XFS_IFEXTIREC) { - xfs_ext_irec_t *erp; /* irec pointer */ - int erp_idx = 0; /* irec index */ - xfs_extnum_t page_idx = idx; /* ext index in target list */ - - erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 0); - return &erp->er_extbuf[page_idx]; - } else if (ifp->if_bytes) { - return &ifp->if_u1.if_extents[idx]; - } else { - return NULL; - } -} - -/* - * Insert new item(s) into the extent records for incore inode - * fork 'ifp'. 'count' new items are inserted at index 'idx'. - */ -void -xfs_iext_insert( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t idx, /* starting index of new items */ - xfs_extnum_t count, /* number of inserted items */ - xfs_bmbt_irec_t *new, /* items to insert */ - int state) /* type of extent conversion */ -{ - xfs_ifork_t *ifp = (state & BMAP_ATTRFORK) ? ip->i_afp : &ip->i_df; - xfs_extnum_t i; /* extent record index */ - - trace_xfs_iext_insert(ip, idx, new, state, _RET_IP_); - - ASSERT(ifp->if_flags & XFS_IFEXTENTS); - xfs_iext_add(ifp, idx, count); - for (i = idx; i < idx + count; i++, new++) - xfs_bmbt_set_all(xfs_iext_get_ext(ifp, i), new); -} + xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn, + &iip->ili_item.li_lsn); -/* - * This is called when the amount of space required for incore file - * extents needs to be increased. The ext_diff parameter stores the - * number of new extents being added and the idx parameter contains - * the extent index where the new extents will be added. If the new - * extents are being appended, then we just need to (re)allocate and - * initialize the space. Otherwise, if the new extents are being - * inserted into the middle of the existing entries, a bit more work - * is required to make room for the new extents to be inserted. The - * caller is responsible for filling in the new extent entries upon - * return. - */ -void -xfs_iext_add( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t idx, /* index to begin adding exts */ - int ext_diff) /* number of extents to add */ -{ - int byte_diff; /* new bytes being added */ - int new_size; /* size of extents after adding */ - xfs_extnum_t nextents; /* number of extents in file */ - - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - ASSERT((idx >= 0) && (idx <= nextents)); - byte_diff = ext_diff * sizeof(xfs_bmbt_rec_t); - new_size = ifp->if_bytes + byte_diff; /* - * If the new number of extents (nextents + ext_diff) - * fits inside the inode, then continue to use the inline - * extent buffer. + * 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. */ - if (nextents + ext_diff <= XFS_INLINE_EXTS) { - if (idx < nextents) { - memmove(&ifp->if_u2.if_inline_ext[idx + ext_diff], - &ifp->if_u2.if_inline_ext[idx], - (nextents - idx) * sizeof(xfs_bmbt_rec_t)); - memset(&ifp->if_u2.if_inline_ext[idx], 0, byte_diff); - } - ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext; - ifp->if_real_bytes = 0; - ifp->if_lastex = nextents + ext_diff; - } - /* - * Otherwise use a linear (direct) extent list. - * If the extents are currently inside the inode, - * xfs_iext_realloc_direct will switch us from - * inline to direct extent allocation mode. - */ - else if (nextents + ext_diff <= XFS_LINEAR_EXTS) { - xfs_iext_realloc_direct(ifp, new_size); - if (idx < nextents) { - memmove(&ifp->if_u1.if_extents[idx + ext_diff], - &ifp->if_u1.if_extents[idx], - (nextents - idx) * sizeof(xfs_bmbt_rec_t)); - memset(&ifp->if_u1.if_extents[idx], 0, byte_diff); - } - } - /* Indirection array */ - else { - xfs_ext_irec_t *erp; - int erp_idx = 0; - int page_idx = idx; - - ASSERT(nextents + ext_diff > XFS_LINEAR_EXTS); - if (ifp->if_flags & XFS_IFEXTIREC) { - erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 1); - } else { - xfs_iext_irec_init(ifp); - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - erp = ifp->if_u1.if_ext_irec; - } - /* Extents fit in target extent page */ - if (erp && erp->er_extcount + ext_diff <= XFS_LINEAR_EXTS) { - if (page_idx < erp->er_extcount) { - memmove(&erp->er_extbuf[page_idx + ext_diff], - &erp->er_extbuf[page_idx], - (erp->er_extcount - page_idx) * - sizeof(xfs_bmbt_rec_t)); - memset(&erp->er_extbuf[page_idx], 0, byte_diff); - } - erp->er_extcount += ext_diff; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff); - } - /* Insert a new extent page */ - else if (erp) { - xfs_iext_add_indirect_multi(ifp, - erp_idx, page_idx, ext_diff); - } - /* - * If extent(s) are being appended to the last page in - * the indirection array and the new extent(s) don't fit - * in the page, then erp is NULL and erp_idx is set to - * the next index needed in the indirection array. - */ - else { - int count = ext_diff; - - while (count) { - erp = xfs_iext_irec_new(ifp, erp_idx); - erp->er_extcount = count; - count -= MIN(count, (int)XFS_LINEAR_EXTS); - if (count) { - erp_idx++; - } - } - } - } - ifp->if_bytes = new_size; -} - -/* - * This is called when incore extents are being added to the indirection - * array and the new extents do not fit in the target extent list. The - * erp_idx parameter contains the irec index for the target extent list - * in the indirection array, and the idx parameter contains the extent - * index within the list. The number of extents being added is stored - * in the count parameter. - * - * |-------| |-------| - * | | | | idx - number of extents before idx - * | idx | | count | - * | | | | count - number of extents being inserted at idx - * |-------| |-------| - * | count | | nex2 | nex2 - number of extents after idx + count - * |-------| |-------| - */ -void -xfs_iext_add_indirect_multi( - xfs_ifork_t *ifp, /* inode fork pointer */ - int erp_idx, /* target extent irec index */ - xfs_extnum_t idx, /* index within target list */ - int count) /* new extents being added */ -{ - int byte_diff; /* new bytes being added */ - xfs_ext_irec_t *erp; /* pointer to irec entry */ - xfs_extnum_t ext_diff; /* number of extents to add */ - xfs_extnum_t ext_cnt; /* new extents still needed */ - xfs_extnum_t nex2; /* extents after idx + count */ - xfs_bmbt_rec_t *nex2_ep = NULL; /* temp list for nex2 extents */ - int nlists; /* number of irec's (lists) */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - nex2 = erp->er_extcount - idx; - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - - /* - * Save second part of target extent list - * (all extents past */ - if (nex2) { - byte_diff = nex2 * sizeof(xfs_bmbt_rec_t); - nex2_ep = (xfs_bmbt_rec_t *) kmem_alloc(byte_diff, KM_NOFS); - memmove(nex2_ep, &erp->er_extbuf[idx], byte_diff); - erp->er_extcount -= nex2; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -nex2); - memset(&erp->er_extbuf[idx], 0, byte_diff); - } + xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item); - /* - * Add the new extents to the end of the target - * list, then allocate new irec record(s) and - * extent buffer(s) as needed to store the rest - * of the new extents. - */ - ext_cnt = count; - ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS - erp->er_extcount); - if (ext_diff) { - erp->er_extcount += ext_diff; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff); - ext_cnt -= ext_diff; - } - while (ext_cnt) { - erp_idx++; - erp = xfs_iext_irec_new(ifp, erp_idx); - ext_diff = MIN(ext_cnt, (int)XFS_LINEAR_EXTS); - erp->er_extcount = ext_diff; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, ext_diff); - ext_cnt -= ext_diff; - } + /* 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); - /* Add nex2 extents back to indirection array */ - if (nex2) { - xfs_extnum_t ext_avail; - int i; + /* generate the checksum. */ + xfs_dinode_calc_crc(mp, dip); - byte_diff = nex2 * sizeof(xfs_bmbt_rec_t); - ext_avail = XFS_LINEAR_EXTS - erp->er_extcount; - i = 0; - /* - * If nex2 extents fit in the current page, append - * nex2_ep after the new extents. - */ - if (nex2 <= ext_avail) { - i = erp->er_extcount; - } - /* - * Otherwise, check if space is available in the - * next page. - */ - else if ((erp_idx < nlists - 1) && - (nex2 <= (ext_avail = XFS_LINEAR_EXTS - - ifp->if_u1.if_ext_irec[erp_idx+1].er_extcount))) { - erp_idx++; - erp++; - /* Create a hole for nex2 extents */ - memmove(&erp->er_extbuf[nex2], erp->er_extbuf, - erp->er_extcount * sizeof(xfs_bmbt_rec_t)); - } - /* - * Final choice, create a new extent page for - * nex2 extents. - */ - else { - erp_idx++; - erp = xfs_iext_irec_new(ifp, erp_idx); - } - memmove(&erp->er_extbuf[i], nex2_ep, byte_diff); - kmem_free(nex2_ep); - erp->er_extcount += nex2; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, nex2); - } -} - -/* - * This is called when the amount of space required for incore file - * extents needs to be decreased. The ext_diff parameter stores the - * number of extents to be removed and the idx parameter contains - * the extent index where the extents will be removed from. - * - * If the amount of space needed has decreased below the linear - * limit, XFS_IEXT_BUFSZ, then switch to using the contiguous - * extent array. Otherwise, use kmem_realloc() to adjust the - * size to what is needed. - */ -void -xfs_iext_remove( - xfs_inode_t *ip, /* incore inode pointer */ - xfs_extnum_t idx, /* index to begin removing exts */ - int ext_diff, /* number of extents to remove */ - int state) /* type of extent conversion */ -{ - xfs_ifork_t *ifp = (state & BMAP_ATTRFORK) ? ip->i_afp : &ip->i_df; - xfs_extnum_t nextents; /* number of extents in file */ - int new_size; /* size of extents after removal */ - - trace_xfs_iext_remove(ip, idx, state, _RET_IP_); - - ASSERT(ext_diff > 0); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - new_size = (nextents - ext_diff) * sizeof(xfs_bmbt_rec_t); - - if (new_size == 0) { - xfs_iext_destroy(ifp); - } else if (ifp->if_flags & XFS_IFEXTIREC) { - xfs_iext_remove_indirect(ifp, idx, ext_diff); - } else if (ifp->if_real_bytes) { - xfs_iext_remove_direct(ifp, idx, ext_diff); - } else { - xfs_iext_remove_inline(ifp, idx, ext_diff); - } - ifp->if_bytes = new_size; -} - -/* - * This removes ext_diff extents from the inline buffer, beginning - * at extent index idx. - */ -void -xfs_iext_remove_inline( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t idx, /* index to begin removing exts */ - int ext_diff) /* number of extents to remove */ -{ - int nextents; /* number of extents in file */ - - ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); - ASSERT(idx < XFS_INLINE_EXTS); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - ASSERT(((nextents - ext_diff) > 0) && - (nextents - ext_diff) < XFS_INLINE_EXTS); - - if (idx + ext_diff < nextents) { - memmove(&ifp->if_u2.if_inline_ext[idx], - &ifp->if_u2.if_inline_ext[idx + ext_diff], - (nextents - (idx + ext_diff)) * - sizeof(xfs_bmbt_rec_t)); - memset(&ifp->if_u2.if_inline_ext[nextents - ext_diff], - 0, ext_diff * sizeof(xfs_bmbt_rec_t)); - } else { - memset(&ifp->if_u2.if_inline_ext[idx], 0, - ext_diff * sizeof(xfs_bmbt_rec_t)); - } -} - -/* - * This removes ext_diff extents from a linear (direct) extent list, - * beginning at extent index idx. If the extents are being removed - * from the end of the list (ie. truncate) then we just need to re- - * allocate the list to remove the extra space. Otherwise, if the - * extents are being removed from the middle of the existing extent - * entries, then we first need to move the extent records beginning - * at idx + ext_diff up in the list to overwrite the records being - * removed, then remove the extra space via kmem_realloc. - */ -void -xfs_iext_remove_direct( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t idx, /* index to begin removing exts */ - int ext_diff) /* number of extents to remove */ -{ - xfs_extnum_t nextents; /* number of extents in file */ - int new_size; /* size of extents after removal */ - - ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); - new_size = ifp->if_bytes - - (ext_diff * sizeof(xfs_bmbt_rec_t)); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - - if (new_size == 0) { - xfs_iext_destroy(ifp); - return; - } - /* Move extents up in the list (if needed) */ - if (idx + ext_diff < nextents) { - memmove(&ifp->if_u1.if_extents[idx], - &ifp->if_u1.if_extents[idx + ext_diff], - (nextents - (idx + ext_diff)) * - sizeof(xfs_bmbt_rec_t)); - } - memset(&ifp->if_u1.if_extents[nextents - ext_diff], - 0, ext_diff * sizeof(xfs_bmbt_rec_t)); - /* - * Reallocate the direct extent list. If the extents - * will fit inside the inode then xfs_iext_realloc_direct - * will switch from direct to inline extent allocation - * mode for us. - */ - xfs_iext_realloc_direct(ifp, new_size); - ifp->if_bytes = new_size; -} - -/* - * This is called when incore extents are being removed from the - * indirection array and the extents being removed span multiple extent - * buffers. The idx parameter contains the file extent index where we - * want to begin removing extents, and the count parameter contains - * how many extents need to be removed. - * - * |-------| |-------| - * | nex1 | | | nex1 - number of extents before idx - * |-------| | count | - * | | | | count - number of extents being removed at idx - * | count | |-------| - * | | | nex2 | nex2 - number of extents after idx + count - * |-------| |-------| - */ -void -xfs_iext_remove_indirect( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t idx, /* index to begin removing extents */ - int count) /* number of extents to remove */ -{ - xfs_ext_irec_t *erp; /* indirection array pointer */ - int erp_idx = 0; /* indirection array index */ - xfs_extnum_t ext_cnt; /* extents left to remove */ - xfs_extnum_t ext_diff; /* extents to remove in current list */ - xfs_extnum_t nex1; /* number of extents before idx */ - xfs_extnum_t nex2; /* extents after idx + count */ - int page_idx = idx; /* index in target extent list */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - erp = xfs_iext_idx_to_irec(ifp, &page_idx, &erp_idx, 0); - ASSERT(erp != NULL); - nex1 = page_idx; - ext_cnt = count; - while (ext_cnt) { - nex2 = MAX((erp->er_extcount - (nex1 + ext_cnt)), 0); - ext_diff = MIN(ext_cnt, (erp->er_extcount - nex1)); - /* - * Check for deletion of entire list; - * xfs_iext_irec_remove() updates extent offsets. - */ - if (ext_diff == erp->er_extcount) { - xfs_iext_irec_remove(ifp, erp_idx); - ext_cnt -= ext_diff; - nex1 = 0; - if (ext_cnt) { - ASSERT(erp_idx < ifp->if_real_bytes / - XFS_IEXT_BUFSZ); - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - nex1 = 0; - continue; - } else { - break; - } - } - /* Move extents up (if needed) */ - if (nex2) { - memmove(&erp->er_extbuf[nex1], - &erp->er_extbuf[nex1 + ext_diff], - nex2 * sizeof(xfs_bmbt_rec_t)); - } - /* Zero out rest of page */ - memset(&erp->er_extbuf[nex1 + nex2], 0, (XFS_IEXT_BUFSZ - - ((nex1 + nex2) * sizeof(xfs_bmbt_rec_t)))); - /* Update remaining counters */ - erp->er_extcount -= ext_diff; - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, -ext_diff); - ext_cnt -= ext_diff; - nex1 = 0; - erp_idx++; - erp++; - } - ifp->if_bytes -= count * sizeof(xfs_bmbt_rec_t); - xfs_iext_irec_compact(ifp); -} - -/* - * Create, destroy, or resize a linear (direct) block of extents. - */ -void -xfs_iext_realloc_direct( - xfs_ifork_t *ifp, /* inode fork pointer */ - int new_size) /* new size of extents */ -{ - int rnew_size; /* real new size of extents */ - - rnew_size = new_size; - - ASSERT(!(ifp->if_flags & XFS_IFEXTIREC) || - ((new_size >= 0) && (new_size <= XFS_IEXT_BUFSZ) && - (new_size != ifp->if_real_bytes))); - - /* Free extent records */ - if (new_size == 0) { - xfs_iext_destroy(ifp); - } - /* Resize direct extent list and zero any new bytes */ - else if (ifp->if_real_bytes) { - /* Check if extents will fit inside the inode */ - if (new_size <= XFS_INLINE_EXTS * sizeof(xfs_bmbt_rec_t)) { - xfs_iext_direct_to_inline(ifp, new_size / - (uint)sizeof(xfs_bmbt_rec_t)); - ifp->if_bytes = new_size; - return; - } - if (!is_power_of_2(new_size)){ - rnew_size = roundup_pow_of_two(new_size); - } - if (rnew_size != ifp->if_real_bytes) { - ifp->if_u1.if_extents = - kmem_realloc(ifp->if_u1.if_extents, - rnew_size, - ifp->if_real_bytes, KM_NOFS); - } - if (rnew_size > ifp->if_real_bytes) { - memset(&ifp->if_u1.if_extents[ifp->if_bytes / - (uint)sizeof(xfs_bmbt_rec_t)], 0, - rnew_size - ifp->if_real_bytes); - } - } - /* - * Switch from the inline extent buffer to a direct - * extent list. Be sure to include the inline extent - * bytes in new_size. - */ - else { - new_size += ifp->if_bytes; - if (!is_power_of_2(new_size)) { - rnew_size = roundup_pow_of_two(new_size); - } - xfs_iext_inline_to_direct(ifp, rnew_size); - } - ifp->if_real_bytes = rnew_size; - ifp->if_bytes = new_size; -} - -/* - * Switch from linear (direct) extent records to inline buffer. - */ -void -xfs_iext_direct_to_inline( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t nextents) /* number of extents in file */ -{ - ASSERT(ifp->if_flags & XFS_IFEXTENTS); - ASSERT(nextents <= XFS_INLINE_EXTS); - /* - * The inline buffer was zeroed when we switched - * from inline to direct extent allocation mode, - * so we don't need to clear it here. - */ - memcpy(ifp->if_u2.if_inline_ext, ifp->if_u1.if_extents, - nextents * sizeof(xfs_bmbt_rec_t)); - kmem_free(ifp->if_u1.if_extents); - ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext; - ifp->if_real_bytes = 0; -} - -/* - * Switch from inline buffer to linear (direct) extent records. - * new_size should already be rounded up to the next power of 2 - * by the caller (when appropriate), so use new_size as it is. - * However, since new_size may be rounded up, we can't update - * if_bytes here. It is the caller's responsibility to update - * if_bytes upon return. - */ -void -xfs_iext_inline_to_direct( - xfs_ifork_t *ifp, /* inode fork pointer */ - int new_size) /* number of extents in file */ -{ - ifp->if_u1.if_extents = kmem_alloc(new_size, KM_NOFS); - memset(ifp->if_u1.if_extents, 0, new_size); - if (ifp->if_bytes) { - memcpy(ifp->if_u1.if_extents, ifp->if_u2.if_inline_ext, - ifp->if_bytes); - memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS * - sizeof(xfs_bmbt_rec_t)); - } - ifp->if_real_bytes = new_size; -} - -/* - * Resize an extent indirection array to new_size bytes. - */ -STATIC void -xfs_iext_realloc_indirect( - xfs_ifork_t *ifp, /* inode fork pointer */ - int new_size) /* new indirection array size */ -{ - int nlists; /* number of irec's (ex lists) */ - int size; /* current indirection array size */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - size = nlists * sizeof(xfs_ext_irec_t); - ASSERT(ifp->if_real_bytes); - ASSERT((new_size >= 0) && (new_size != size)); - if (new_size == 0) { - xfs_iext_destroy(ifp); - } else { - ifp->if_u1.if_ext_irec = (xfs_ext_irec_t *) - kmem_realloc(ifp->if_u1.if_ext_irec, - new_size, size, KM_NOFS); - } -} - -/* - * Switch from indirection array to linear (direct) extent allocations. - */ -STATIC void -xfs_iext_indirect_to_direct( - xfs_ifork_t *ifp) /* inode fork pointer */ -{ - xfs_bmbt_rec_host_t *ep; /* extent record pointer */ - xfs_extnum_t nextents; /* number of extents in file */ - int size; /* size of file extents */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - ASSERT(nextents <= XFS_LINEAR_EXTS); - size = nextents * sizeof(xfs_bmbt_rec_t); - - xfs_iext_irec_compact_pages(ifp); - ASSERT(ifp->if_real_bytes == XFS_IEXT_BUFSZ); - - ep = ifp->if_u1.if_ext_irec->er_extbuf; - kmem_free(ifp->if_u1.if_ext_irec); - ifp->if_flags &= ~XFS_IFEXTIREC; - ifp->if_u1.if_extents = ep; - ifp->if_bytes = size; - if (nextents < XFS_LINEAR_EXTS) { - xfs_iext_realloc_direct(ifp, size); - } -} - -/* - * Free incore file extents. - */ -void -xfs_iext_destroy( - xfs_ifork_t *ifp) /* inode fork pointer */ -{ - if (ifp->if_flags & XFS_IFEXTIREC) { - int erp_idx; - int nlists; - - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - for (erp_idx = nlists - 1; erp_idx >= 0 ; erp_idx--) { - xfs_iext_irec_remove(ifp, erp_idx); - } - ifp->if_flags &= ~XFS_IFEXTIREC; - } else if (ifp->if_real_bytes) { - kmem_free(ifp->if_u1.if_extents); - } else if (ifp->if_bytes) { - memset(ifp->if_u2.if_inline_ext, 0, XFS_INLINE_EXTS * - sizeof(xfs_bmbt_rec_t)); - } - ifp->if_u1.if_extents = NULL; - ifp->if_real_bytes = 0; - ifp->if_bytes = 0; -} - -/* - * Return a pointer to the extent record for file system block bno. - */ -xfs_bmbt_rec_host_t * /* pointer to found extent record */ -xfs_iext_bno_to_ext( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_fileoff_t bno, /* block number to search for */ - xfs_extnum_t *idxp) /* index of target extent */ -{ - xfs_bmbt_rec_host_t *base; /* pointer to first extent */ - xfs_filblks_t blockcount = 0; /* number of blocks in extent */ - xfs_bmbt_rec_host_t *ep = NULL; /* pointer to target extent */ - xfs_ext_irec_t *erp = NULL; /* indirection array pointer */ - int high; /* upper boundary in search */ - xfs_extnum_t idx = 0; /* index of target extent */ - int low; /* lower boundary in search */ - xfs_extnum_t nextents; /* number of file extents */ - xfs_fileoff_t startoff = 0; /* start offset of extent */ - - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - if (nextents == 0) { - *idxp = 0; - return NULL; - } - low = 0; - if (ifp->if_flags & XFS_IFEXTIREC) { - /* Find target extent list */ - int erp_idx = 0; - erp = xfs_iext_bno_to_irec(ifp, bno, &erp_idx); - base = erp->er_extbuf; - high = erp->er_extcount - 1; - } else { - base = ifp->if_u1.if_extents; - high = nextents - 1; - } - /* Binary search extent records */ - while (low <= high) { - idx = (low + high) >> 1; - ep = base + idx; - startoff = xfs_bmbt_get_startoff(ep); - blockcount = xfs_bmbt_get_blockcount(ep); - if (bno < startoff) { - high = idx - 1; - } else if (bno >= startoff + blockcount) { - low = idx + 1; - } else { - /* Convert back to file-based extent index */ - if (ifp->if_flags & XFS_IFEXTIREC) { - idx += erp->er_extoff; - } - *idxp = idx; - return ep; - } - } - /* Convert back to file-based extent index */ - if (ifp->if_flags & XFS_IFEXTIREC) { - idx += erp->er_extoff; - } - if (bno >= startoff + blockcount) { - if (++idx == nextents) { - ep = NULL; - } else { - ep = xfs_iext_get_ext(ifp, idx); - } - } - *idxp = idx; - return ep; -} - -/* - * Return a pointer to the indirection array entry containing the - * extent record for filesystem block bno. Store the index of the - * target irec in *erp_idxp. - */ -xfs_ext_irec_t * /* pointer to found extent record */ -xfs_iext_bno_to_irec( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_fileoff_t bno, /* block number to search for */ - int *erp_idxp) /* irec index of target ext list */ -{ - xfs_ext_irec_t *erp = NULL; /* indirection array pointer */ - xfs_ext_irec_t *erp_next; /* next indirection array entry */ - int erp_idx; /* indirection array index */ - int nlists; /* number of extent irec's (lists) */ - int high; /* binary search upper limit */ - int low; /* binary search lower limit */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - erp_idx = 0; - low = 0; - high = nlists - 1; - while (low <= high) { - erp_idx = (low + high) >> 1; - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - erp_next = erp_idx < nlists - 1 ? erp + 1 : NULL; - if (bno < xfs_bmbt_get_startoff(erp->er_extbuf)) { - high = erp_idx - 1; - } else if (erp_next && bno >= - xfs_bmbt_get_startoff(erp_next->er_extbuf)) { - low = erp_idx + 1; - } else { - break; - } - } - *erp_idxp = erp_idx; - return erp; -} - -/* - * Return a pointer to the indirection array entry containing the - * extent record at file extent index *idxp. Store the index of the - * target irec in *erp_idxp and store the page index of the target - * extent record in *idxp. - */ -xfs_ext_irec_t * -xfs_iext_idx_to_irec( - xfs_ifork_t *ifp, /* inode fork pointer */ - xfs_extnum_t *idxp, /* extent index (file -> page) */ - int *erp_idxp, /* pointer to target irec */ - int realloc) /* new bytes were just added */ -{ - xfs_ext_irec_t *prev; /* pointer to previous irec */ - xfs_ext_irec_t *erp = NULL; /* pointer to current irec */ - int erp_idx; /* indirection array index */ - int nlists; /* number of irec's (ex lists) */ - int high; /* binary search upper limit */ - int low; /* binary search lower limit */ - xfs_extnum_t page_idx = *idxp; /* extent index in target list */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - ASSERT(page_idx >= 0 && page_idx <= - ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t)); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - erp_idx = 0; - low = 0; - high = nlists - 1; - - /* Binary search extent irec's */ - while (low <= high) { - erp_idx = (low + high) >> 1; - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - prev = erp_idx > 0 ? erp - 1 : NULL; - if (page_idx < erp->er_extoff || (page_idx == erp->er_extoff && - realloc && prev && prev->er_extcount < XFS_LINEAR_EXTS)) { - high = erp_idx - 1; - } else if (page_idx > erp->er_extoff + erp->er_extcount || - (page_idx == erp->er_extoff + erp->er_extcount && - !realloc)) { - low = erp_idx + 1; - } else if (page_idx == erp->er_extoff + erp->er_extcount && - erp->er_extcount == XFS_LINEAR_EXTS) { - ASSERT(realloc); - page_idx = 0; - erp_idx++; - erp = erp_idx < nlists ? erp + 1 : NULL; - break; - } else { - page_idx -= erp->er_extoff; - break; - } - } - *idxp = page_idx; - *erp_idxp = erp_idx; - return(erp); -} - -/* - * Allocate and initialize an indirection array once the space needed - * for incore extents increases above XFS_IEXT_BUFSZ. - */ -void -xfs_iext_irec_init( - xfs_ifork_t *ifp) /* inode fork pointer */ -{ - xfs_ext_irec_t *erp; /* indirection array pointer */ - xfs_extnum_t nextents; /* number of extents in file */ - - ASSERT(!(ifp->if_flags & XFS_IFEXTIREC)); - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - ASSERT(nextents <= XFS_LINEAR_EXTS); - - erp = kmem_alloc(sizeof(xfs_ext_irec_t), KM_NOFS); - - if (nextents == 0) { - ifp->if_u1.if_extents = kmem_alloc(XFS_IEXT_BUFSZ, KM_NOFS); - } else if (!ifp->if_real_bytes) { - xfs_iext_inline_to_direct(ifp, XFS_IEXT_BUFSZ); - } else if (ifp->if_real_bytes < XFS_IEXT_BUFSZ) { - xfs_iext_realloc_direct(ifp, XFS_IEXT_BUFSZ); - } - erp->er_extbuf = ifp->if_u1.if_extents; - erp->er_extcount = nextents; - erp->er_extoff = 0; - - ifp->if_flags |= XFS_IFEXTIREC; - ifp->if_real_bytes = XFS_IEXT_BUFSZ; - ifp->if_bytes = nextents * sizeof(xfs_bmbt_rec_t); - ifp->if_u1.if_ext_irec = erp; - - return; -} - -/* - * Allocate and initialize a new entry in the indirection array. - */ -xfs_ext_irec_t * -xfs_iext_irec_new( - xfs_ifork_t *ifp, /* inode fork pointer */ - int erp_idx) /* index for new irec */ -{ - xfs_ext_irec_t *erp; /* indirection array pointer */ - int i; /* loop counter */ - int nlists; /* number of irec's (ex lists) */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - - /* Resize indirection array */ - xfs_iext_realloc_indirect(ifp, ++nlists * - sizeof(xfs_ext_irec_t)); - /* - * Move records down in the array so the - * new page can use erp_idx. - */ - erp = ifp->if_u1.if_ext_irec; - for (i = nlists - 1; i > erp_idx; i--) { - memmove(&erp[i], &erp[i-1], sizeof(xfs_ext_irec_t)); - } - ASSERT(i == erp_idx); - - /* Initialize new extent record */ - erp = ifp->if_u1.if_ext_irec; - erp[erp_idx].er_extbuf = kmem_alloc(XFS_IEXT_BUFSZ, KM_NOFS); - ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ; - memset(erp[erp_idx].er_extbuf, 0, XFS_IEXT_BUFSZ); - erp[erp_idx].er_extcount = 0; - erp[erp_idx].er_extoff = erp_idx > 0 ? - erp[erp_idx-1].er_extoff + erp[erp_idx-1].er_extcount : 0; - return (&erp[erp_idx]); -} - -/* - * Remove a record from the indirection array. - */ -void -xfs_iext_irec_remove( - xfs_ifork_t *ifp, /* inode fork pointer */ - int erp_idx) /* irec index to remove */ -{ - xfs_ext_irec_t *erp; /* indirection array pointer */ - int i; /* loop counter */ - int nlists; /* number of irec's (ex lists) */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - if (erp->er_extbuf) { - xfs_iext_irec_update_extoffs(ifp, erp_idx + 1, - -erp->er_extcount); - kmem_free(erp->er_extbuf); - } - /* Compact extent records */ - erp = ifp->if_u1.if_ext_irec; - for (i = erp_idx; i < nlists - 1; i++) { - memmove(&erp[i], &erp[i+1], sizeof(xfs_ext_irec_t)); - } - /* - * Manually free the last extent record from the indirection - * array. A call to xfs_iext_realloc_indirect() with a size - * of zero would result in a call to xfs_iext_destroy() which - * would in turn call this function again, creating a nasty - * infinite loop. - */ - if (--nlists) { - xfs_iext_realloc_indirect(ifp, - nlists * sizeof(xfs_ext_irec_t)); - } else { - kmem_free(ifp->if_u1.if_ext_irec); - } - ifp->if_real_bytes = nlists * XFS_IEXT_BUFSZ; -} - -/* - * This is called to clean up large amounts of unused memory allocated - * by the indirection array. Before compacting anything though, verify - * that the indirection array is still needed and switch back to the - * linear extent list (or even the inline buffer) if possible. The - * compaction policy is as follows: - * - * Full Compaction: Extents fit into a single page (or inline buffer) - * Partial Compaction: Extents occupy less than 50% of allocated space - * No Compaction: Extents occupy at least 50% of allocated space - */ -void -xfs_iext_irec_compact( - xfs_ifork_t *ifp) /* inode fork pointer */ -{ - xfs_extnum_t nextents; /* number of extents in file */ - int nlists; /* number of irec's (ex lists) */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - nextents = ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t); - - if (nextents == 0) { - xfs_iext_destroy(ifp); - } else if (nextents <= XFS_INLINE_EXTS) { - xfs_iext_indirect_to_direct(ifp); - xfs_iext_direct_to_inline(ifp, nextents); - } else if (nextents <= XFS_LINEAR_EXTS) { - xfs_iext_indirect_to_direct(ifp); - } else if (nextents < (nlists * XFS_LINEAR_EXTS) >> 1) { - xfs_iext_irec_compact_pages(ifp); - } -} - -/* - * Combine extents from neighboring extent pages. - */ -void -xfs_iext_irec_compact_pages( - xfs_ifork_t *ifp) /* inode fork pointer */ -{ - xfs_ext_irec_t *erp, *erp_next;/* pointers to irec entries */ - int erp_idx = 0; /* indirection array index */ - int nlists; /* number of irec's (ex lists) */ - - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - while (erp_idx < nlists - 1) { - erp = &ifp->if_u1.if_ext_irec[erp_idx]; - erp_next = erp + 1; - if (erp_next->er_extcount <= - (XFS_LINEAR_EXTS - erp->er_extcount)) { - memcpy(&erp->er_extbuf[erp->er_extcount], - erp_next->er_extbuf, erp_next->er_extcount * - sizeof(xfs_bmbt_rec_t)); - erp->er_extcount += erp_next->er_extcount; - /* - * Free page before removing extent record - * so er_extoffs don't get modified in - * xfs_iext_irec_remove. - */ - kmem_free(erp_next->er_extbuf); - erp_next->er_extbuf = NULL; - xfs_iext_irec_remove(ifp, erp_idx + 1); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - } else { - erp_idx++; - } - } -} - -/* - * This is called to update the er_extoff field in the indirection - * array when extents have been added or removed from one of the - * extent lists. erp_idx contains the irec index to begin updating - * at and ext_diff contains the number of extents that were added - * or removed. - */ -void -xfs_iext_irec_update_extoffs( - xfs_ifork_t *ifp, /* inode fork pointer */ - int erp_idx, /* irec index to update */ - int ext_diff) /* number of new extents */ -{ - int i; /* loop counter */ - int nlists; /* number of irec's (ex lists */ + ASSERT(bp->b_fspriv != NULL); + ASSERT(bp->b_iodone != NULL); + return 0; - ASSERT(ifp->if_flags & XFS_IFEXTIREC); - nlists = ifp->if_real_bytes / XFS_IEXT_BUFSZ; - for (i = erp_idx; i < nlists; i++) { - ifp->if_u1.if_ext_irec[i].er_extoff += ext_diff; - } +corrupt_out: + return XFS_ERROR(EFSCORRUPTED); } |