| /linux/fs/xfs/scrub/ |
| H A D | attr_repair.c | 515 * multiblock buffers once we drop the ILOCK. 561 * The caller must hold the ILOCK on the file being repaired. We use 821 * sc->ip's ILOCK to prevent rename from updating parent pointer in xrep_xattr_full_reset() 822 * information and the tempfile's ILOCK to clear its contents. in xrep_xattr_full_reset() 851 * should no longer be joined to the transaction, so we drop its ILOCK. in xrep_xattr_full_reset() 1003 * fork. The caller must ILOCK the scrub file and join it to the transaction. 1030 * fork. The caller must ILOCK the tempfile and join it to the transaction. 1082 * the ILOCK independently of the transaction. in xrep_xattr_salvage_attributes() 1392 * Repair relies on the ILOCK to quiesce all possible xattr updates. in xrep_xattr_finalize_tempfile() 1492 * can drop the tempfile's ILOCK an in xrep_xattr_rebuild_tree() [all...] |
| H A D | dirtree.c | 50 * subdirectory, so we instead synchronize on ILOCK and use dirent update hooks 269 * parent pointer in @sc->ip. We probably dropped @sc->ip's ILOCK while 573 * that we can acquire and drop the ILOCK on an alias of sc->ip. The in xchk_dirpath_walk_upwards() 574 * ILOCK state is no longer tracked in the scrub context. Hence we in xchk_dirpath_walk_upwards() 575 * must drop @sc->ip's ILOCK during the walk. in xchk_dirpath_walk_upwards() 965 * directory's ILOCK, which means that any in-progress directory update in xchk_dirtree()
|
| H A D | dirtree_repair.c | 422 * the ILOCK. in xrep_dirtree_unlink() 529 * drop the scan lock, the ILOCK, and the transaction so that in xrep_dirtree_delete_path() 812 * ILOCK -> scan lock. in xrep_dirtree()
|
| H A D | parent.c | 260 /* If we've cycled the ILOCK, we must revalidate all deferred pptrs. */ 596 * If we can grab both IOLOCK and ILOCK of the alleged parent, we in xchk_parent_slow_pptr() 710 * If we cycled the ILOCK while cross-checking parent pointers with in xchk_parent_count_pptrs()
|
| H A D | dqiterate.c | 166 * zero before the first call, and must not hold the quota file ILOCK.
|
| H A D | quota.c | 157 * operations, the locking order is first the ILOCK and then the dquot. in xchk_quota_item()
|
| H A D | nlinks_repair.c | 161 * We still hold @ip's ILOCK, so its link count cannot change. in xrep_nlinks_repair_inode()
|
| H A D | nlinks.c | 97 * observations is inode ILOCK -> iscan_lock/xchk_nlink_ctrs lock. 1020 * already scanned, and the scanner thread takes each inode's ILOCK, in xchk_nlinks_setup_scan()
|
| H A D | rtbitmap_repair.c | 48 * because we cannot drop the rtbitmap ILOCK) and cannot ask for more
|
| H A D | cow_repair.c | 490 * seriously wrong, since we didn't drop the ILOCK.
|
| H A D | inode_repair.c | 304 * AGI buffer, we cannot block waiting for the ILOCK because rename can take 305 * the ILOCK and then try to lock AGIs.
|
| H A D | bmap.c | 557 * on disk (we still hold the ILOCK) and skip the equivalence check. in xchk_bmapbt_rec()
|
| /linux/fs/xfs/ |
| H A D | xfs_inode_item.h | 29 * ILOCK and IFLUSHING and hence we need some other method of
|
| H A D | xfs_iops.c | 42 * buffer inside filldir(), and this happens with the ilock on the directory 44 * mmap_lock is taken during the page fault, and then we lock the ilock to do 45 * block mapping. Hence we need a different class for the directory ilock so 802 * If the IDs do change before we take the ilock, we're covered in xfs_setattr_nonsize()
|
| H A D | xfs_inode.c | 115 * the mmap lock second and the ilock last in order to prevent deadlock. 296 * Sometimes we assert the ILOCK is held exclusively, but we're in in xfs_assert_ilocked() 377 * the iolock, the mmaplock or the ilock, but not more than one at a time. If we 477 * xfs_lock_two_inodes() can only be used to lock ilock. The iolock and 2581 * that, and grab the ilock while still holding the i_flags_lock in xfs_iflush_cluster() 2593 * ILOCK will pin the inode against reclaim and prevent in xfs_iflush_cluster()
|
| H A D | xfs_reflink.c | 470 * The data fork mapping may have changed while we dropped the ILOCK in xfs_reflink_fill_cow_hole() 545 * ILOCK (a racing O_DIRECT writer under IOLOCK_SHARED can in xfs_reflink_fill_delalloc() 985 * repeatedly cycles the ILOCK to allocate one transaction per remapped in xfs_reflink_end_cow()
|
| H A D | xfs_dquot.c | 989 * allocate blocks if needed. The inode's ILOCK must be held and it must not 1037 * dropped the ilock. in xfs_qm_dqget_inode()
|
| H A D | xfs_aops.c | 189 * It's safe to check the COW fork if_seq here without the ILOCK because in xfs_map_blocks()
|
| H A D | xfs_icache.c | 1003 * We do this as early as possible under the ILOCK so that in xfs_reclaim_inode() 1626 * Callers must not hold any inode's ILOCK. If requesting a synchronous scan
|
| H A D | xfs_ioctl.c | 794 * If the IDs do change before we take the ilock, we're covered in xfs_fileattr_set()
|
| H A D | xfs_zone_alloc.c | 430 * All dereferences of rtg->rtg_open_zone hold the ILOCK for the rmap in xfs_init_open_zone()
|
| H A D | xfs_qm.c | 1336 * We don't _need_ to take the ilock EXCL here because quotacheck runs in xfs_qm_dqusage_adjust()
|
| H A D | xfs_iomap.c | 920 /* may drop and re-acquire the ilock */ in xfs_direct_write_iomap_begin()
|