^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * Copyright (c) 2000-2006 Silicon Graphics, Inc.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * All Rights Reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include <linux/iversion.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include "xfs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include "xfs_fs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include "xfs_shared.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include "xfs_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include "xfs_log_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include "xfs_trans_resv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include "xfs_sb.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include "xfs_mount.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include "xfs_defer.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include "xfs_inode.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include "xfs_dir2.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include "xfs_attr.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include "xfs_trans_space.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include "xfs_trans.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include "xfs_buf_item.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include "xfs_inode_item.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include "xfs_ialloc.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include "xfs_bmap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include "xfs_bmap_util.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include "xfs_errortag.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include "xfs_error.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include "xfs_quota.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include "xfs_filestream.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include "xfs_trace.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) #include "xfs_icache.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #include "xfs_symlink.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) #include "xfs_trans_priv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) #include "xfs_log.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) #include "xfs_bmap_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) #include "xfs_reflink.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) kmem_zone_t *xfs_inode_zone;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) * Used in xfs_itruncate_extents(). This is the maximum number of extents
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) * freed from a file in a single transaction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) #define XFS_ITRUNC_MAX_EXTENTS 2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) STATIC int xfs_iunlink(struct xfs_trans *, struct xfs_inode *);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) STATIC int xfs_iunlink_remove(struct xfs_trans *, struct xfs_inode *);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) * helper function to extract extent size hint from inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) xfs_extlen_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) xfs_get_extsz_hint(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) * No point in aligning allocations if we need to COW to actually
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * write to them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) if (xfs_is_always_cow_inode(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) if ((ip->i_d.di_flags & XFS_DIFLAG_EXTSIZE) && ip->i_d.di_extsize)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) return ip->i_d.di_extsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) if (XFS_IS_REALTIME_INODE(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) return ip->i_mount->m_sb.sb_rextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) * Helper function to extract CoW extent size hint from inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * Between the extent size hint and the CoW extent size hint, we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * return the greater of the two. If the value is zero (automatic),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) * use the default size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) xfs_extlen_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) xfs_get_cowextsz_hint(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) xfs_extlen_t a, b;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) a = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) if (ip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) a = ip->i_d.di_cowextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) b = xfs_get_extsz_hint(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) a = max(a, b);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) if (a == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) return XFS_DEFAULT_COWEXTSZ_HINT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) return a;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) * These two are wrapper routines around the xfs_ilock() routine used to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) * centralize some grungy code. They are used in places that wish to lock the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * inode solely for reading the extents. The reason these places can't just
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) * call xfs_ilock(ip, XFS_ILOCK_SHARED) is that the inode lock also guards to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * bringing in of the extents from disk for a file in b-tree format. If the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) * inode is in b-tree format, then we need to lock the inode exclusively until
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) * the extents are read in. Locking it exclusively all the time would limit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) * our parallelism unnecessarily, though. What we do instead is check to see
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) * if the extents have been read in yet, and only lock the inode exclusively
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) * if they have not.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) * The functions return a value which should be given to the corresponding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) * xfs_iunlock() call.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) uint
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) xfs_ilock_data_map_shared(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) uint lock_mode = XFS_ILOCK_SHARED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) if (ip->i_df.if_format == XFS_DINODE_FMT_BTREE &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) (ip->i_df.if_flags & XFS_IFEXTENTS) == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) lock_mode = XFS_ILOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) xfs_ilock(ip, lock_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) return lock_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) uint
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) xfs_ilock_attr_map_shared(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) uint lock_mode = XFS_ILOCK_SHARED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) if (ip->i_afp &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) ip->i_afp->if_format == XFS_DINODE_FMT_BTREE &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) lock_mode = XFS_ILOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) xfs_ilock(ip, lock_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) return lock_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) * In addition to i_rwsem in the VFS inode, the xfs inode contains 2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) * multi-reader locks: i_mmap_lock and the i_lock. This routine allows
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) * various combinations of the locks to be obtained.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) * The 3 locks should always be ordered so that the IO lock is obtained first,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) * the mmap lock second and the ilock last in order to prevent deadlock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) * Basic locking order:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) * i_rwsem -> i_mmap_lock -> page_lock -> i_ilock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) * mmap_lock locking order:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) * i_rwsem -> page lock -> mmap_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) * mmap_lock -> i_mmap_lock -> page_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) * The difference in mmap_lock locking order mean that we cannot hold the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) * i_mmap_lock over syscall based read(2)/write(2) based IO. These IO paths can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) * fault in pages during copy in/out (for buffered IO) or require the mmap_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) * in get_user_pages() to map the user pages into the kernel address space for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) * direct IO. Similarly the i_rwsem cannot be taken inside a page fault because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) * page faults already hold the mmap_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * Hence to serialise fully against both syscall and mmap based IO, we need to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * take both the i_rwsem and the i_mmap_lock. These locks should *only* be both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) * taken in places where we need to invalidate the page cache in a race
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) * free manner (e.g. truncate, hole punch and other extent manipulation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) * functions).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) xfs_ilock(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) xfs_inode_t *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) uint lock_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) trace_xfs_ilock(ip, lock_flags, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) * You can't set both SHARED and EXCL for the same lock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) if (lock_flags & XFS_IOLOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) down_write_nested(&VFS_I(ip)->i_rwsem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) XFS_IOLOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) } else if (lock_flags & XFS_IOLOCK_SHARED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) down_read_nested(&VFS_I(ip)->i_rwsem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) XFS_IOLOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) if (lock_flags & XFS_MMAPLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) mrupdate_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) else if (lock_flags & XFS_MMAPLOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) mraccess_nested(&ip->i_mmaplock, XFS_MMAPLOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) if (lock_flags & XFS_ILOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) mrupdate_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) else if (lock_flags & XFS_ILOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) mraccess_nested(&ip->i_lock, XFS_ILOCK_DEP(lock_flags));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) * This is just like xfs_ilock(), except that the caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) * is guaranteed not to sleep. It returns 1 if it gets
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * the requested locks and 0 otherwise. If the IO lock is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) * obtained but the inode lock cannot be, then the IO lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) * is dropped before returning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) * ip -- the inode being locked
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) * lock_flags -- this parameter indicates the inode's locks to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) * to be locked. See the comment for xfs_ilock() for a list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) * of valid values.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) xfs_ilock_nowait(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) xfs_inode_t *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) uint lock_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) trace_xfs_ilock_nowait(ip, lock_flags, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) * You can't set both SHARED and EXCL for the same lock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) if (lock_flags & XFS_IOLOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) if (!down_write_trylock(&VFS_I(ip)->i_rwsem))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) } else if (lock_flags & XFS_IOLOCK_SHARED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) if (!down_read_trylock(&VFS_I(ip)->i_rwsem))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) if (lock_flags & XFS_MMAPLOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) if (!mrtryupdate(&ip->i_mmaplock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) goto out_undo_iolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) } else if (lock_flags & XFS_MMAPLOCK_SHARED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) if (!mrtryaccess(&ip->i_mmaplock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) goto out_undo_iolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) if (lock_flags & XFS_ILOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) if (!mrtryupdate(&ip->i_lock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) goto out_undo_mmaplock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) } else if (lock_flags & XFS_ILOCK_SHARED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) if (!mrtryaccess(&ip->i_lock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) goto out_undo_mmaplock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) out_undo_mmaplock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) if (lock_flags & XFS_MMAPLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) mrunlock_excl(&ip->i_mmaplock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) else if (lock_flags & XFS_MMAPLOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) mrunlock_shared(&ip->i_mmaplock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) out_undo_iolock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) if (lock_flags & XFS_IOLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) up_write(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) else if (lock_flags & XFS_IOLOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) up_read(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) * xfs_iunlock() is used to drop the inode locks acquired with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) * xfs_ilock() and xfs_ilock_nowait(). The caller must pass
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) * in the flags given to xfs_ilock() or xfs_ilock_nowait() so
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) * that we know which locks to drop.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) * ip -- the inode being unlocked
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) * lock_flags -- this parameter indicates the inode's locks to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) * to be unlocked. See the comment for xfs_ilock() for a list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) * of valid values for this parameter.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) xfs_iunlock(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) xfs_inode_t *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) uint lock_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) * You can't set both SHARED and EXCL for the same lock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) * and only XFS_IOLOCK_SHARED, XFS_IOLOCK_EXCL, XFS_ILOCK_SHARED,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) * and XFS_ILOCK_EXCL are valid values to set in lock_flags.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) ASSERT((lock_flags & (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) (XFS_IOLOCK_SHARED | XFS_IOLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) ASSERT((lock_flags & (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) (XFS_MMAPLOCK_SHARED | XFS_MMAPLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) ASSERT((lock_flags & (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL)) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) (XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) ASSERT((lock_flags & ~(XFS_LOCK_MASK | XFS_LOCK_SUBCLASS_MASK)) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) ASSERT(lock_flags != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) if (lock_flags & XFS_IOLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) up_write(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) else if (lock_flags & XFS_IOLOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) up_read(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) if (lock_flags & XFS_MMAPLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) mrunlock_excl(&ip->i_mmaplock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) else if (lock_flags & XFS_MMAPLOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) mrunlock_shared(&ip->i_mmaplock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) if (lock_flags & XFS_ILOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) mrunlock_excl(&ip->i_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) else if (lock_flags & XFS_ILOCK_SHARED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) mrunlock_shared(&ip->i_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) trace_xfs_iunlock(ip, lock_flags, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) * give up write locks. the i/o lock cannot be held nested
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) * if it is being demoted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) xfs_ilock_demote(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) xfs_inode_t *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) uint lock_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) ASSERT(lock_flags & (XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) ASSERT((lock_flags &
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) ~(XFS_IOLOCK_EXCL|XFS_MMAPLOCK_EXCL|XFS_ILOCK_EXCL)) == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) if (lock_flags & XFS_ILOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) mrdemote(&ip->i_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) if (lock_flags & XFS_MMAPLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) mrdemote(&ip->i_mmaplock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) if (lock_flags & XFS_IOLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) downgrade_write(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) trace_xfs_ilock_demote(ip, lock_flags, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) #if defined(DEBUG) || defined(XFS_WARN)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) xfs_isilocked(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) xfs_inode_t *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) uint lock_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) if (!(lock_flags & XFS_ILOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) return !!ip->i_lock.mr_writer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) return rwsem_is_locked(&ip->i_lock.mr_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) if (lock_flags & (XFS_MMAPLOCK_EXCL|XFS_MMAPLOCK_SHARED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) if (!(lock_flags & XFS_MMAPLOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) return !!ip->i_mmaplock.mr_writer;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) return rwsem_is_locked(&ip->i_mmaplock.mr_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) if (!(lock_flags & XFS_IOLOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) return !debug_locks ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) lockdep_is_held_type(&VFS_I(ip)->i_rwsem, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) return rwsem_is_locked(&VFS_I(ip)->i_rwsem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) ASSERT(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) * xfs_lockdep_subclass_ok() is only used in an ASSERT, so is only called when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) * DEBUG or XFS_WARN is set. And MAX_LOCKDEP_SUBCLASSES is then only defined
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) * when CONFIG_LOCKDEP is set. Hence the complex define below to avoid build
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) * errors and warnings.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) #if (defined(DEBUG) || defined(XFS_WARN)) && defined(CONFIG_LOCKDEP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) static bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) xfs_lockdep_subclass_ok(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) return subclass < MAX_LOCKDEP_SUBCLASSES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) #define xfs_lockdep_subclass_ok(subclass) (true)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) * Bump the subclass so xfs_lock_inodes() acquires each lock with a different
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) * value. This can be called for any type of inode lock combination, including
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) * parent locking. Care must be taken to ensure we don't overrun the subclass
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) * storage fields in the class mask we build.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400) static inline int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) xfs_lock_inumorder(int lock_mode, int subclass)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) int class = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) ASSERT(!(lock_mode & (XFS_ILOCK_PARENT | XFS_ILOCK_RTBITMAP |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) XFS_ILOCK_RTSUM)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) ASSERT(xfs_lockdep_subclass_ok(subclass));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) if (lock_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) ASSERT(subclass <= XFS_IOLOCK_MAX_SUBCLASS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) class += subclass << XFS_IOLOCK_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) if (lock_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) ASSERT(subclass <= XFS_MMAPLOCK_MAX_SUBCLASS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) class += subclass << XFS_MMAPLOCK_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) if (lock_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420) ASSERT(subclass <= XFS_ILOCK_MAX_SUBCLASS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) class += subclass << XFS_ILOCK_SHIFT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) return (lock_mode & ~XFS_LOCK_SUBCLASS_MASK) | class;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) * The following routine will lock n inodes in exclusive mode. We assume the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) * caller calls us with the inodes in i_ino order.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) * We need to detect deadlock where an inode that we lock is in the AIL and we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) * start waiting for another inode that is locked by a thread in a long running
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) * transaction (such as truncate). This can result in deadlock since the long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) * running trans might need to wait for the inode we just locked in order to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) * push the tail and free space in the log.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) * xfs_lock_inodes() can only be used to lock one type of lock at a time -
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) * the iolock, the mmaplock or the ilock, but not more than one at a time. If we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) * lock more than one at a time, lockdep will report false positives saying we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) * have violated locking orders.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) xfs_lock_inodes(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) struct xfs_inode **ips,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) int inodes,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) uint lock_mode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) int attempts = 0, i, j, try_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) struct xfs_log_item *lp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) * Currently supports between 2 and 5 inodes with exclusive locking. We
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) * support an arbitrary depth of locking here, but absolute limits on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * inodes depend on the type of locking and the limits placed by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) * lockdep annotations in xfs_lock_inumorder. These are all checked by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) * the asserts.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) ASSERT(ips && inodes >= 2 && inodes <= 5);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) ASSERT(lock_mode & (XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) ASSERT(!(lock_mode & (XFS_IOLOCK_SHARED | XFS_MMAPLOCK_SHARED |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) XFS_ILOCK_SHARED)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) ASSERT(!(lock_mode & XFS_MMAPLOCK_EXCL) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) inodes <= XFS_MMAPLOCK_MAX_SUBCLASS + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) ASSERT(!(lock_mode & XFS_ILOCK_EXCL) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) inodes <= XFS_ILOCK_MAX_SUBCLASS + 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) if (lock_mode & XFS_IOLOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) ASSERT(!(lock_mode & (XFS_MMAPLOCK_EXCL | XFS_ILOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) } else if (lock_mode & XFS_MMAPLOCK_EXCL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) ASSERT(!(lock_mode & XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) try_lock = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) again:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) for (; i < inodes; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) ASSERT(ips[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) if (i && (ips[i] == ips[i - 1])) /* Already locked */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) * If try_lock is not set yet, make sure all locked inodes are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) * not in the AIL. If any are, set try_lock to be used later.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) if (!try_lock) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) for (j = (i - 1); j >= 0 && !try_lock; j--) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) lp = &ips[j]->i_itemp->ili_item;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) if (lp && test_bit(XFS_LI_IN_AIL, &lp->li_flags))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) try_lock++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) * If any of the previous locks we have locked is in the AIL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) * we must TRY to get the second and subsequent locks. If
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) * we can't get any, we must release all we have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) * and try again.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) if (!try_lock) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) xfs_ilock(ips[i], xfs_lock_inumorder(lock_mode, i));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) /* try_lock means we have an inode locked that is in the AIL. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) ASSERT(i != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) if (xfs_ilock_nowait(ips[i], xfs_lock_inumorder(lock_mode, i)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) * Unlock all previous guys and try again. xfs_iunlock will try
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) * to push the tail if the inode is in the AIL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) attempts++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) for (j = i - 1; j >= 0; j--) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) * Check to see if we've already unlocked this one. Not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) * the first one going back, and the inode ptr is the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) * same.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) if (j != (i - 1) && ips[j] == ips[j + 1])
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) xfs_iunlock(ips[j], lock_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) if ((attempts % 5) == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) delay(1); /* Don't just spin the CPU */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) try_lock = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) goto again;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) * xfs_lock_two_inodes() can only be used to lock one type of lock at a time -
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) * the mmaplock or the ilock, but not more than one type at a time. If we lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) * more than one at a time, lockdep will report false positives saying we have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) * violated locking orders. The iolock must be double-locked separately since
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) * we use i_rwsem for that. We now support taking one lock EXCL and the other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) * SHARED.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) xfs_lock_two_inodes(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) struct xfs_inode *ip0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) uint ip0_mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) struct xfs_inode *ip1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) uint ip1_mode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) struct xfs_inode *temp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) uint mode_temp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) int attempts = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) struct xfs_log_item *lp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) ASSERT(hweight32(ip0_mode) == 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) ASSERT(hweight32(ip1_mode) == 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) ASSERT(!(ip0_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) ASSERT(!(ip1_mode & (XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) ASSERT(!(ip0_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) !(ip0_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) ASSERT(!(ip1_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) !(ip1_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) ASSERT(!(ip1_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) !(ip0_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) ASSERT(!(ip0_mode & (XFS_MMAPLOCK_SHARED|XFS_MMAPLOCK_EXCL)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) !(ip1_mode & (XFS_ILOCK_SHARED|XFS_ILOCK_EXCL)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) ASSERT(ip0->i_ino != ip1->i_ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) if (ip0->i_ino > ip1->i_ino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) temp = ip0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) ip0 = ip1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) ip1 = temp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) mode_temp = ip0_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) ip0_mode = ip1_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) ip1_mode = mode_temp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) again:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) xfs_ilock(ip0, xfs_lock_inumorder(ip0_mode, 0));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) * If the first lock we have locked is in the AIL, we must TRY to get
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) * the second lock. If we can't get it, we must release the first one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) * and try again.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) lp = &ip0->i_itemp->ili_item;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) if (lp && test_bit(XFS_LI_IN_AIL, &lp->li_flags)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) if (!xfs_ilock_nowait(ip1, xfs_lock_inumorder(ip1_mode, 1))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) xfs_iunlock(ip0, ip0_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) if ((++attempts % 5) == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) delay(1); /* Don't just spin the CPU */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) goto again;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) xfs_ilock(ip1, xfs_lock_inumorder(ip1_mode, 1));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) STATIC uint
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) _xfs_dic2xflags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) uint16_t di_flags,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) uint64_t di_flags2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) bool has_attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) uint flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) if (di_flags & XFS_DIFLAG_ANY) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) if (di_flags & XFS_DIFLAG_REALTIME)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) flags |= FS_XFLAG_REALTIME;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) if (di_flags & XFS_DIFLAG_PREALLOC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) flags |= FS_XFLAG_PREALLOC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) if (di_flags & XFS_DIFLAG_IMMUTABLE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) flags |= FS_XFLAG_IMMUTABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) if (di_flags & XFS_DIFLAG_APPEND)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) flags |= FS_XFLAG_APPEND;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) if (di_flags & XFS_DIFLAG_SYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) flags |= FS_XFLAG_SYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) if (di_flags & XFS_DIFLAG_NOATIME)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) flags |= FS_XFLAG_NOATIME;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) if (di_flags & XFS_DIFLAG_NODUMP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) flags |= FS_XFLAG_NODUMP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) if (di_flags & XFS_DIFLAG_RTINHERIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) flags |= FS_XFLAG_RTINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) if (di_flags & XFS_DIFLAG_PROJINHERIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) flags |= FS_XFLAG_PROJINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) if (di_flags & XFS_DIFLAG_NOSYMLINKS)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) flags |= FS_XFLAG_NOSYMLINKS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) if (di_flags & XFS_DIFLAG_EXTSIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) flags |= FS_XFLAG_EXTSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) if (di_flags & XFS_DIFLAG_EXTSZINHERIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) flags |= FS_XFLAG_EXTSZINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) if (di_flags & XFS_DIFLAG_NODEFRAG)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) flags |= FS_XFLAG_NODEFRAG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) if (di_flags & XFS_DIFLAG_FILESTREAM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) flags |= FS_XFLAG_FILESTREAM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) if (di_flags2 & XFS_DIFLAG2_ANY) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) if (di_flags2 & XFS_DIFLAG2_DAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) flags |= FS_XFLAG_DAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) if (di_flags2 & XFS_DIFLAG2_COWEXTSIZE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) flags |= FS_XFLAG_COWEXTSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) if (has_attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) flags |= FS_XFLAG_HASATTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) return flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) uint
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) xfs_ip2xflags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) struct xfs_icdinode *dic = &ip->i_d;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) return _xfs_dic2xflags(dic->di_flags, dic->di_flags2, XFS_IFORK_Q(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) * Lookups up an inode from "name". If ci_name is not NULL, then a CI match
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) * is allowed, otherwise it has to be an exact match. If a CI match is found,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) * ci_name->name will point to a the actual name (caller must free) or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) * will be set to NULL if an exact match is found.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) xfs_lookup(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) xfs_inode_t *dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) struct xfs_name *name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) xfs_inode_t **ipp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) struct xfs_name *ci_name)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) xfs_ino_t inum;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) trace_xfs_lookup(dp, name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) if (XFS_FORCED_SHUTDOWN(dp->i_mount))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) error = xfs_dir_lookup(NULL, dp, name, &inum, ci_name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) error = xfs_iget(dp->i_mount, NULL, inum, 0, 0, ipp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) goto out_free_name;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) out_free_name:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) if (ci_name)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) kmem_free(ci_name->name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) /* Propagate di_flags from a parent inode to a child inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) xfs_inode_inherit_flags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) const struct xfs_inode *pip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) unsigned int di_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) umode_t mode = VFS_I(ip)->i_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) if (S_ISDIR(mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) if (pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) di_flags |= XFS_DIFLAG_RTINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) di_flags |= XFS_DIFLAG_EXTSZINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) ip->i_d.di_extsize = pip->i_d.di_extsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) if (pip->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) di_flags |= XFS_DIFLAG_PROJINHERIT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) } else if (S_ISREG(mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) if ((pip->i_d.di_flags & XFS_DIFLAG_RTINHERIT) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) xfs_sb_version_hasrealtime(&ip->i_mount->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) di_flags |= XFS_DIFLAG_REALTIME;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) if (pip->i_d.di_flags & XFS_DIFLAG_EXTSZINHERIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) di_flags |= XFS_DIFLAG_EXTSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) ip->i_d.di_extsize = pip->i_d.di_extsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) if ((pip->i_d.di_flags & XFS_DIFLAG_NOATIME) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) xfs_inherit_noatime)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) di_flags |= XFS_DIFLAG_NOATIME;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) if ((pip->i_d.di_flags & XFS_DIFLAG_NODUMP) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) xfs_inherit_nodump)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) di_flags |= XFS_DIFLAG_NODUMP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) if ((pip->i_d.di_flags & XFS_DIFLAG_SYNC) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) xfs_inherit_sync)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) di_flags |= XFS_DIFLAG_SYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) if ((pip->i_d.di_flags & XFS_DIFLAG_NOSYMLINKS) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) xfs_inherit_nosymlinks)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) di_flags |= XFS_DIFLAG_NOSYMLINKS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) if ((pip->i_d.di_flags & XFS_DIFLAG_NODEFRAG) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) xfs_inherit_nodefrag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) di_flags |= XFS_DIFLAG_NODEFRAG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) if (pip->i_d.di_flags & XFS_DIFLAG_FILESTREAM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) di_flags |= XFS_DIFLAG_FILESTREAM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) ip->i_d.di_flags |= di_flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) /* Propagate di_flags2 from a parent inode to a child inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) xfs_inode_inherit_flags2(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) const struct xfs_inode *pip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) if (pip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) ip->i_d.di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) ip->i_d.di_cowextsize = pip->i_d.di_cowextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) if (pip->i_d.di_flags2 & XFS_DIFLAG2_DAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) ip->i_d.di_flags2 |= XFS_DIFLAG2_DAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) * Allocate an inode on disk and return a copy of its in-core version.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) * The in-core inode is locked exclusively. Set mode, nlink, and rdev
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) * appropriately within the inode. The uid and gid for the inode are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) * set according to the contents of the given cred structure.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) * Use xfs_dialloc() to allocate the on-disk inode. If xfs_dialloc()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) * has a free inode available, call xfs_iget() to obtain the in-core
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) * version of the allocated inode. Finally, fill in the inode and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) * log its initial contents. In this case, ialloc_context would be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) * set to NULL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) * If xfs_dialloc() does not have an available inode, it will replenish
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) * its supply by doing an allocation. Since we can only do one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) * allocation within a transaction without deadlocks, we must commit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) * the current transaction before returning the inode itself.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) * In this case, therefore, we will set ialloc_context and return.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) * The caller should then commit the current transaction, start a new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) * transaction, and call xfs_ialloc() again to actually get the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) * To ensure that some other process does not grab the inode that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) * was allocated during the first call to xfs_ialloc(), this routine
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) * also returns the [locked] bp pointing to the head of the freelist
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) * as ialloc_context. The caller should hold this buffer across
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) * the commit and pass it back into this routine on the second call.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) * If we are allocating quota inodes, we do not have a parent inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) * to attach to or associate with (i.e. pip == NULL) because they
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) * are not linked into the directory structure - they are attached
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) * directly to the superblock - and so have no parent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) xfs_ialloc(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) xfs_trans_t *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) xfs_inode_t *pip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) umode_t mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) xfs_nlink_t nlink,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) dev_t rdev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) prid_t prid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) xfs_buf_t **ialloc_context,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) xfs_inode_t **ipp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) xfs_ino_t ino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) xfs_inode_t *ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) uint flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) struct timespec64 tv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) struct inode *inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) * Call the space management code to pick
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) * the on-disk inode to be allocated.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) error = xfs_dialloc(tp, pip ? pip->i_ino : 0, mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) ialloc_context, &ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) if (*ialloc_context || ino == NULLFSINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) ASSERT(*ialloc_context == NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) * Protect against obviously corrupt allocation btree records. Later
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) * xfs_iget checks will catch re-allocation of other active in-memory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) * and on-disk inodes. If we don't catch reallocating the parent inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) * here we will deadlock in xfs_iget() so we have to do these checks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) * first.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) if ((pip && ino == pip->i_ino) || !xfs_verify_dir_ino(mp, ino)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) xfs_alert(mp, "Allocated a known in-use inode 0x%llx!", ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) return -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) * Get the in-core inode with the lock held exclusively.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) * This is because we're setting fields here we need
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) * to prevent others from looking at until we're done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) error = xfs_iget(mp, tp, ino, XFS_IGET_CREATE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) XFS_ILOCK_EXCL, &ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) ASSERT(ip != NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) inode = VFS_I(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) inode->i_mode = mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) set_nlink(inode, nlink);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) inode->i_uid = current_fsuid();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) inode->i_rdev = rdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) ip->i_d.di_projid = prid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) if (pip && XFS_INHERIT_GID(pip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) inode->i_gid = VFS_I(pip)->i_gid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) if ((VFS_I(pip)->i_mode & S_ISGID) && S_ISDIR(mode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) inode->i_mode |= S_ISGID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) inode->i_gid = current_fsgid();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) * If the group ID of the new file does not match the effective group
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) * ID or one of the supplementary group IDs, the S_ISGID bit is cleared
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) * (and only if the irix_sgid_inherit compatibility variable is set).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) if (irix_sgid_inherit &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) (inode->i_mode & S_ISGID) && !in_group_p(inode->i_gid))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) inode->i_mode &= ~S_ISGID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) ip->i_d.di_size = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) ip->i_df.if_nextents = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) ASSERT(ip->i_d.di_nblocks == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) tv = current_time(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) inode->i_mtime = tv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) inode->i_atime = tv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) inode->i_ctime = tv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) ip->i_d.di_extsize = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) ip->i_d.di_dmevmask = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) ip->i_d.di_dmstate = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) ip->i_d.di_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) inode_set_iversion(inode, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) ip->i_d.di_flags2 = mp->m_ino_geo.new_diflags2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890) ip->i_d.di_cowextsize = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) ip->i_d.di_crtime = tv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) flags = XFS_ILOG_CORE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) switch (mode & S_IFMT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) case S_IFIFO:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) case S_IFCHR:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) case S_IFBLK:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) case S_IFSOCK:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) ip->i_df.if_format = XFS_DINODE_FMT_DEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) ip->i_df.if_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902) flags |= XFS_ILOG_DEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) case S_IFREG:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) case S_IFDIR:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) if (pip && (pip->i_d.di_flags & XFS_DIFLAG_ANY))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) xfs_inode_inherit_flags(ip, pip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) if (pip && (pip->i_d.di_flags2 & XFS_DIFLAG2_ANY))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) xfs_inode_inherit_flags2(ip, pip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) /* FALLTHROUGH */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) case S_IFLNK:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912) ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913) ip->i_df.if_flags = XFS_IFEXTENTS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) ip->i_df.if_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) ip->i_df.if_u1.if_root = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) ASSERT(0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) * Log the new values stuffed into the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) xfs_trans_log_inode(tp, ip, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) /* now that we have an i_mode we can setup the inode structure */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) xfs_setup_inode(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) *ipp = ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) * Allocates a new inode from disk and return a pointer to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) * incore copy. This routine will internally commit the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) * transaction and allocate a new one if the Space Manager needed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) * to do an allocation to replenish the inode free-list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) * This routine is designed to be called from xfs_create and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) * xfs_create_dir.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) xfs_dir_ialloc(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) xfs_trans_t **tpp, /* input: current transaction;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) output: may be a new transaction. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) xfs_inode_t *dp, /* directory within whose allocate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) the inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) umode_t mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) xfs_nlink_t nlink,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) dev_t rdev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) prid_t prid, /* project id */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) xfs_inode_t **ipp) /* pointer to inode; it will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) locked. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) xfs_trans_t *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) xfs_inode_t *ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) xfs_buf_t *ialloc_context = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960) int code;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) void *dqinfo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) uint tflags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) tp = *tpp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) * xfs_ialloc will return a pointer to an incore inode if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) * the Space Manager has an available inode on the free
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) * list. Otherwise, it will do an allocation and replenish
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) * the freelist. Since we can only do one allocation per
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) * transaction without deadlocks, we will need to commit the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) * current transaction and start a new one. We will then
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) * need to call xfs_ialloc again to get the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) * If xfs_ialloc did an allocation to replenish the freelist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) * it returns the bp containing the head of the freelist as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) * ialloc_context. We will hold a lock on it across the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) * transaction commit so that no other process can steal
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) * the inode(s) that we've just allocated.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid, &ialloc_context,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) &ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986) * Return an error if we were unable to allocate a new inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) * This should only happen if we run out of space on disk or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) * encounter a disk error.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) if (code) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) return code;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) if (!ialloc_context && !ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) return -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) * If the AGI buffer is non-NULL, then we were unable to get an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) * inode in one operation. We need to commit the current
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) * transaction and call xfs_ialloc() again. It is guaranteed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) * to succeed the second time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) if (ialloc_context) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) * Normally, xfs_trans_commit releases all the locks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) * We call bhold to hang on to the ialloc_context across
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) * the commit. Holding this buffer prevents any other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) * processes from doing any allocations in this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) * allocation group.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) xfs_trans_bhold(tp, ialloc_context);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) * We want the quota changes to be associated with the next
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) * transaction, NOT this one. So, detach the dqinfo from this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) * and attach it to the next transaction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) dqinfo = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) tflags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) if (tp->t_dqinfo) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) dqinfo = (void *)tp->t_dqinfo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) tp->t_dqinfo = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) tflags = tp->t_flags & XFS_TRANS_DQ_DIRTY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) tp->t_flags &= ~(XFS_TRANS_DQ_DIRTY);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) code = xfs_trans_roll(&tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) * Re-attach the quota info that we detached from prev trx.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) if (dqinfo) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) tp->t_dqinfo = dqinfo;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) tp->t_flags |= tflags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) if (code) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) xfs_buf_relse(ialloc_context);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) *tpp = tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) return code;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) xfs_trans_bjoin(tp, ialloc_context);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) * Call ialloc again. Since we've locked out all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049) * other allocations in this allocation group,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) * this call should always succeed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) code = xfs_ialloc(tp, dp, mode, nlink, rdev, prid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) &ialloc_context, &ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) * If we get an error at this point, return to the caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) * so that the current transaction can be aborted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059) if (code) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) *tpp = tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061) *ipp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) return code;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) ASSERT(!ialloc_context && ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) *ipp = ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) *tpp = tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) * Decrement the link count on an inode & log the change. If this causes the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076) * link count to go to zero, move the inode to AGI unlinked list so that it can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) * be freed when the last active reference goes away via xfs_inactive().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079) static int /* error */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) xfs_droplink(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) xfs_trans_t *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) xfs_inode_t *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) drop_nlink(VFS_I(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) if (VFS_I(ip)->i_nlink)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) return xfs_iunlink(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096) * Increment the link count on an inode & log the change.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) xfs_bumplink(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) xfs_trans_t *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) xfs_inode_t *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) inc_nlink(VFS_I(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) xfs_create(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) xfs_inode_t *dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) struct xfs_name *name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) umode_t mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) dev_t rdev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) xfs_inode_t **ipp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) int is_dir = S_ISDIR(mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) struct xfs_mount *mp = dp->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) struct xfs_inode *ip = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) struct xfs_trans *tp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) bool unlock_dp_on_error = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) prid_t prid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) struct xfs_dquot *udqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) struct xfs_dquot *gdqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) struct xfs_dquot *pdqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) struct xfs_trans_res *tres;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) uint resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) trace_xfs_create(dp, name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) prid = xfs_get_initial_prid(dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) * Make sure that we have allocated dquot(s) on disk.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141) XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) &udqp, &gdqp, &pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) if (is_dir) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147) resblks = XFS_MKDIR_SPACE_RES(mp, name->len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148) tres = &M_RES(mp)->tr_mkdir;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) resblks = XFS_CREATE_SPACE_RES(mp, name->len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) tres = &M_RES(mp)->tr_create;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155) * Initially assume that the file does not exist and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) * reserve the resources for that case. If that is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) * the case we'll drop the one we have and get a more
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) * appropriate transaction later.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) if (error == -ENOSPC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) /* flush outstanding delalloc blocks and retry */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) xfs_flush_inodes(mp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) goto out_release_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) xfs_ilock(dp, XFS_ILOCK_EXCL | XFS_ILOCK_PARENT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) unlock_dp_on_error = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173) * Reserve disk quota and the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175) error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) pdqp, resblks, 1, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181) * A newly created regular or special file just has one directory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) * entry pointing to them, but a directory also the "." entry
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) * pointing to itself.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185) error = xfs_dir_ialloc(&tp, dp, mode, is_dir ? 2 : 1, rdev, prid, &ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) * Now we join the directory inode to the transaction. We do not do it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) * earlier because xfs_dir_ialloc might commit the previous transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) * (and release all the locks). An error from here on will result in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) * the transaction cancel unlocking dp so don't do it explicitly in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) * error path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) unlock_dp_on_error = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) error = xfs_dir_createname(tp, dp, name, ip->i_ino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200) resblks - XFS_IALLOC_SPACE_RES(mp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) ASSERT(error != -ENOSPC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206) xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208) if (is_dir) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209) error = xfs_dir_init(tp, ip, dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) xfs_bumplink(tp, dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217) * If this is a synchronous mount, make sure that the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218) * create transaction goes to disk before returning to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219) * the user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1223)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1224) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1225) * Attach the dquot(s) to the inodes and modify them incore.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226) * These ids of the inode couldn't have changed since the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227) * inode has been locked ever since it was created.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229) xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) goto out_release_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235) xfs_qm_dqrele(udqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) xfs_qm_dqrele(gdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) xfs_qm_dqrele(pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239) *ipp = ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) out_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244) out_release_inode:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246) * Wait until after the current transaction is aborted to finish the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247) * setup of the inode and release the inode. This prevents recursive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) * transactions and deadlocks from xfs_inactive.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250) if (ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251) xfs_finish_inode_setup(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252) xfs_irele(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) xfs_qm_dqrele(udqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) xfs_qm_dqrele(gdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) xfs_qm_dqrele(pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259) if (unlock_dp_on_error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) xfs_iunlock(dp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265) xfs_create_tmpfile(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) struct xfs_inode *dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267) umode_t mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) struct xfs_inode **ipp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270) struct xfs_mount *mp = dp->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271) struct xfs_inode *ip = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) struct xfs_trans *tp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) prid_t prid;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) struct xfs_dquot *udqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) struct xfs_dquot *gdqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277) struct xfs_dquot *pdqp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278) struct xfs_trans_res *tres;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279) uint resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) prid = xfs_get_initial_prid(dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) * Make sure that we have allocated dquot(s) on disk.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291) &udqp, &gdqp, &pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295) resblks = XFS_IALLOC_SPACE_RES(mp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296) tres = &M_RES(mp)->tr_create_tmpfile;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) error = xfs_trans_alloc(mp, tres, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300) goto out_release_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302) error = xfs_trans_reserve_quota(tp, mp, udqp, gdqp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) pdqp, resblks, 1, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) error = xfs_dir_ialloc(&tp, dp, mode, 0, 0, prid, &ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) if (mp->m_flags & XFS_MOUNT_WSYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315) * Attach the dquot(s) to the inodes and modify them incore.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316) * These ids of the inode couldn't have changed since the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) * inode has been locked ever since it was created.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1321) error = xfs_iunlink(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1322) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) goto out_release_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) xfs_qm_dqrele(udqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330) xfs_qm_dqrele(gdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) xfs_qm_dqrele(pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333) *ipp = ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336) out_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338) out_release_inode:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) * Wait until after the current transaction is aborted to finish the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341) * setup of the inode and release the inode. This prevents recursive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) * transactions and deadlocks from xfs_inactive.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344) if (ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) xfs_finish_inode_setup(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) xfs_irele(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) xfs_qm_dqrele(udqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350) xfs_qm_dqrele(gdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) xfs_qm_dqrele(pdqp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1352)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1353) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1354) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) xfs_link(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) xfs_inode_t *tdp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359) xfs_inode_t *sip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360) struct xfs_name *target_name)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) xfs_mount_t *mp = tdp->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) xfs_trans_t *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365) int resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) trace_xfs_link(tdp, target_name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369) ASSERT(!S_ISDIR(VFS_I(sip)->i_mode));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) error = xfs_qm_dqattach(sip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378) error = xfs_qm_dqattach(tdp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) resblks = XFS_LINK_SPACE_RES(mp, target_name->len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384) if (error == -ENOSPC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385) resblks = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_link, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) xfs_lock_two_inodes(sip, XFS_ILOCK_EXCL, tdp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397) * If we are using project inheritance, we only allow hard link
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) * creation in our tree when the project IDs are the same; else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) * the tree quota mechanism could be circumvented.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1400) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1401) if (unlikely((tdp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1402) tdp->i_d.di_projid != sip->i_d.di_projid)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1403) error = -EXDEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1404) goto error_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1405) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1406)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1407) if (!resblks) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1408) error = xfs_dir_canenter(tp, tdp, target_name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1409) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1410) goto error_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1411) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1412)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1413) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1414) * Handle initial link state of O_TMPFILE inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1415) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1416) if (VFS_I(sip)->i_nlink == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1417) error = xfs_iunlink_remove(tp, sip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1418) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1419) goto error_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1420) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1422) error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1423) resblks);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1424) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1425) goto error_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1426) xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1427) xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1428)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1429) xfs_bumplink(tp, sip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1430)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1431) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1432) * If this is a synchronous mount, make sure that the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1433) * link transaction goes to disk before returning to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1434) * the user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1435) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1436) if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1437) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1438)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1439) return xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1441) error_return:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1442) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1443) std_return:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1444) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1445) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1446)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1447) /* Clear the reflink flag and the cowblocks tag if possible. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1448) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1449) xfs_itruncate_clear_reflink_flags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1450) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1451) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1452) struct xfs_ifork *dfork;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1453) struct xfs_ifork *cfork;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1454)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1455) if (!xfs_is_reflink_inode(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1456) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1457) dfork = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1458) cfork = XFS_IFORK_PTR(ip, XFS_COW_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1459) if (dfork->if_bytes == 0 && cfork->if_bytes == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1460) ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1461) if (cfork->if_bytes == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1462) xfs_inode_clear_cowblocks_tag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1463) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1464)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1465) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1466) * Free up the underlying blocks past new_size. The new size must be smaller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1467) * than the current size. This routine can be used both for the attribute and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1468) * data fork, and does not modify the inode size, which is left to the caller.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1469) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1470) * The transaction passed to this routine must have made a permanent log
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1471) * reservation of at least XFS_ITRUNCATE_LOG_RES. This routine may commit the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1472) * given transaction and start new ones, so make sure everything involved in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1473) * the transaction is tidy before calling here. Some transaction will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1474) * returned to the caller to be committed. The incoming transaction must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1475) * already include the inode, and both inode locks must be held exclusively.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1476) * The inode must also be "held" within the transaction. On return the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1477) * will be "held" within the returned transaction. This routine does NOT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1478) * require any disk space to be reserved for it within the transaction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1479) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1480) * If we get an error, we must return with the inode locked and linked into the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1481) * current transaction. This keeps things simple for the higher level code,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1482) * because it always knows that the inode is locked and held in the transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1483) * that returns to it whether errors occur or not. We don't mark the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1484) * dirty on error so that transactions can be easily aborted if possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1485) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1486) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1487) xfs_itruncate_extents_flags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1488) struct xfs_trans **tpp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1489) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1490) int whichfork,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1491) xfs_fsize_t new_size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1492) int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1493) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1494) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1495) struct xfs_trans *tp = *tpp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1496) xfs_fileoff_t first_unmap_block;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1497) xfs_filblks_t unmap_len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1498) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1499)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1500) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1501) ASSERT(!atomic_read(&VFS_I(ip)->i_count) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1502) xfs_isilocked(ip, XFS_IOLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1503) ASSERT(new_size <= XFS_ISIZE(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1504) ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1505) ASSERT(ip->i_itemp != NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1506) ASSERT(ip->i_itemp->ili_lock_flags == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1507) ASSERT(!XFS_NOT_DQATTACHED(mp, ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1508)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1509) trace_xfs_itruncate_extents_start(ip, new_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1510)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1511) flags |= xfs_bmapi_aflag(whichfork);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1512)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1513) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1514) * Since it is possible for space to become allocated beyond
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1515) * the end of the file (in a crash where the space is allocated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1516) * but the inode size is not yet updated), simply remove any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1517) * blocks which show up between the new EOF and the maximum
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1518) * possible file size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1519) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1520) * We have to free all the blocks to the bmbt maximum offset, even if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1521) * the page cache can't scale that far.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1522) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1523) first_unmap_block = XFS_B_TO_FSB(mp, (xfs_ufsize_t)new_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1524) if (first_unmap_block >= XFS_MAX_FILEOFF) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1525) WARN_ON_ONCE(first_unmap_block > XFS_MAX_FILEOFF);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1526) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1527) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1528)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1529) unmap_len = XFS_MAX_FILEOFF - first_unmap_block + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1530) while (unmap_len > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1531) ASSERT(tp->t_firstblock == NULLFSBLOCK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1532) error = __xfs_bunmapi(tp, ip, first_unmap_block, &unmap_len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1533) flags, XFS_ITRUNC_MAX_EXTENTS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1534) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1535) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1537) /* free the just unmapped extents */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1538) error = xfs_defer_finish(&tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1539) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1540) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1541) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1542)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1543) if (whichfork == XFS_DATA_FORK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1544) /* Remove all pending CoW reservations. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1545) error = xfs_reflink_cancel_cow_blocks(ip, &tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1546) first_unmap_block, XFS_MAX_FILEOFF, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1547) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1548) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1549)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1550) xfs_itruncate_clear_reflink_flags(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1551) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1553) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1554) * Always re-log the inode so that our permanent transaction can keep
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1555) * on rolling it forward in the log.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1556) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1557) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1558)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1559) trace_xfs_itruncate_extents_end(ip, new_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1560)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1561) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1562) *tpp = tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1563) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1564) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1566) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1567) xfs_release(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1568) xfs_inode_t *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1569) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1570) xfs_mount_t *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1571) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1573) if (!S_ISREG(VFS_I(ip)->i_mode) || (VFS_I(ip)->i_mode == 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1574) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1575)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1576) /* If this is a read-only mount, don't do this (would generate I/O) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1577) if (mp->m_flags & XFS_MOUNT_RDONLY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1578) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1580) if (!XFS_FORCED_SHUTDOWN(mp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1581) int truncated;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1582)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1583) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1584) * If we previously truncated this file and removed old data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1585) * in the process, we want to initiate "early" writeout on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1586) * the last close. This is an attempt to combat the notorious
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1587) * NULL files problem which is particularly noticeable from a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1588) * truncate down, buffered (re-)write (delalloc), followed by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1589) * a crash. What we are effectively doing here is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1590) * significantly reducing the time window where we'd otherwise
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1591) * be exposed to that problem.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1592) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1593) truncated = xfs_iflags_test_and_clear(ip, XFS_ITRUNCATED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1594) if (truncated) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1595) xfs_iflags_clear(ip, XFS_IDIRTY_RELEASE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1596) if (ip->i_delayed_blks > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1597) error = filemap_flush(VFS_I(ip)->i_mapping);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1598) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1599) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1600) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1601) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1602) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1603)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1604) if (VFS_I(ip)->i_nlink == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1605) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1606)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1607) if (xfs_can_free_eofblocks(ip, false)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1608)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1609) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1610) * Check if the inode is being opened, written and closed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1611) * frequently and we have delayed allocation blocks outstanding
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1612) * (e.g. streaming writes from the NFS server), truncating the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1613) * blocks past EOF will cause fragmentation to occur.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1614) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1615) * In this case don't do the truncation, but we have to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1616) * careful how we detect this case. Blocks beyond EOF show up as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1617) * i_delayed_blks even when the inode is clean, so we need to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1618) * truncate them away first before checking for a dirty release.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1619) * Hence on the first dirty close we will still remove the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1620) * speculative allocation, but after that we will leave it in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1621) * place.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1622) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1623) if (xfs_iflags_test(ip, XFS_IDIRTY_RELEASE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1624) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1625) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1626) * If we can't get the iolock just skip truncating the blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1627) * past EOF because we could deadlock with the mmap_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1628) * otherwise. We'll get another chance to drop them once the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1629) * last reference to the inode is dropped, so we'll never leak
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1630) * blocks permanently.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1631) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1632) if (xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1633) error = xfs_free_eofblocks(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1634) xfs_iunlock(ip, XFS_IOLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1635) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1636) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1637) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1638)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1639) /* delalloc blocks after truncation means it really is dirty */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1640) if (ip->i_delayed_blks)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1641) xfs_iflags_set(ip, XFS_IDIRTY_RELEASE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1642) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1643) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1644) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1645)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1646) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1647) * xfs_inactive_truncate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1648) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1649) * Called to perform a truncate when an inode becomes unlinked.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1650) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1651) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1652) xfs_inactive_truncate(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1653) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1654) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1655) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1656) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1657) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1658)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1659) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1660) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1661) ASSERT(XFS_FORCED_SHUTDOWN(mp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1662) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1663) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1664) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1665) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1666)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1667) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1668) * Log the inode size first to prevent stale data exposure in the event
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1669) * of a system crash before the truncate completes. See the related
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1670) * comment in xfs_vn_setattr_size() for details.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1671) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1672) ip->i_d.di_size = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1673) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1674)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1675) error = xfs_itruncate_extents(&tp, ip, XFS_DATA_FORK, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1676) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1677) goto error_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1678)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1679) ASSERT(ip->i_df.if_nextents == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1680)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1681) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1682) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1683) goto error_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1684)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1685) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1686) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1687)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1688) error_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1689) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1690) error_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1691) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1692) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1693) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1694)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1695) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1696) * xfs_inactive_ifree()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1697) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1698) * Perform the inode free when an inode is unlinked.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1699) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1700) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1701) xfs_inactive_ifree(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1702) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1703) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1704) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1705) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1706) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1707)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1708) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1709) * We try to use a per-AG reservation for any block needed by the finobt
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1710) * tree, but as the finobt feature predates the per-AG reservation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1711) * support a degraded file system might not have enough space for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1712) * reservation at mount time. In that case try to dip into the reserved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1713) * pool and pray.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1714) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1715) * Send a warning if the reservation does happen to fail, as the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1716) * now remains allocated and sits on the unlinked list until the fs is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1717) * repaired.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1718) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1719) if (unlikely(mp->m_finobt_nores)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1720) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1721) XFS_IFREE_SPACE_RES(mp), 0, XFS_TRANS_RESERVE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1722) &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1723) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1724) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ifree, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1725) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1726) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1727) if (error == -ENOSPC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1728) xfs_warn_ratelimited(mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1729) "Failed to remove inode(s) from unlinked list. "
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1730) "Please free space, unmount and run xfs_repair.");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1731) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1732) ASSERT(XFS_FORCED_SHUTDOWN(mp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1733) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1734) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1735) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1736)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1737) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1738) * We do not hold the inode locked across the entire rolling transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1739) * here. We only need to hold it for the first transaction that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1740) * xfs_ifree() builds, which may mark the inode XFS_ISTALE if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1741) * underlying cluster buffer is freed. Relogging an XFS_ISTALE inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1742) * here breaks the relationship between cluster buffer invalidation and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1743) * stale inode invalidation on cluster buffer item journal commit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1744) * completion, and can result in leaving dirty stale inodes hanging
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1745) * around in memory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1746) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1747) * We have no need for serialising this inode operation against other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1748) * operations - we freed the inode and hence reallocation is required
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1749) * and that will serialise on reallocating the space the deferops need
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1750) * to free. Hence we can unlock the inode on the first commit of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1751) * the transaction rather than roll it right through the deferops. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1752) * avoids relogging the XFS_ISTALE inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1753) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1754) * We check that xfs_ifree() hasn't grown an internal transaction roll
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1755) * by asserting that the inode is still locked when it returns.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1756) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1757) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1758) xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1759)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1760) error = xfs_ifree(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1761) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1762) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1763) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1764) * If we fail to free the inode, shut down. The cancel
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1765) * might do that, we need to make sure. Otherwise the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1766) * inode might be lost for a long time or forever.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1767) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1768) if (!XFS_FORCED_SHUTDOWN(mp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1769) xfs_notice(mp, "%s: xfs_ifree returned error %d",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1770) __func__, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1771) xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1772) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1773) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1774) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1775) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1776)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1777) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1778) * Credit the quota account(s). The inode is gone.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1779) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1780) xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_ICOUNT, -1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1781)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1782) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1783) * Just ignore errors at this point. There is nothing we can do except
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1784) * to try to keep going. Make sure it's not a silent error.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1785) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1786) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1787) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1788) xfs_notice(mp, "%s: xfs_trans_commit returned error %d",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1789) __func__, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1791) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1792) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1793)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1794) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1795) * xfs_inactive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1796) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1797) * This is called when the vnode reference count for the vnode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1798) * goes to zero. If the file has been unlinked, then it must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1799) * now be truncated. Also, we clear all of the read-ahead state
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1800) * kept for the inode here since the file is now closed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1801) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1802) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1803) xfs_inactive(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1804) xfs_inode_t *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1805) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1806) struct xfs_mount *mp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1807) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1808) int truncate = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1809)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1810) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1811) * If the inode is already free, then there can be nothing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1812) * to clean up here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1813) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1814) if (VFS_I(ip)->i_mode == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1815) ASSERT(ip->i_df.if_broot_bytes == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1816) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1817) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1818)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1819) mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1820) ASSERT(!xfs_iflags_test(ip, XFS_IRECOVERY));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1821)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1822) /* If this is a read-only mount, don't do this (would generate I/O) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1823) if (mp->m_flags & XFS_MOUNT_RDONLY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1824) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1825)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1826) /* Try to clean out the cow blocks if there are any. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1827) if (xfs_inode_has_cow_data(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1828) xfs_reflink_cancel_cow_range(ip, 0, NULLFILEOFF, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1829)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1830) if (VFS_I(ip)->i_nlink != 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1831) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1832) * force is true because we are evicting an inode from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1833) * cache. Post-eof blocks must be freed, lest we end up with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1834) * broken free space accounting.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1835) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1836) * Note: don't bother with iolock here since lockdep complains
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1837) * about acquiring it in reclaim context. We have the only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1838) * reference to the inode at this point anyways.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1839) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1840) if (xfs_can_free_eofblocks(ip, true))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1841) xfs_free_eofblocks(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1842)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1843) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1844) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1845)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1846) if (S_ISREG(VFS_I(ip)->i_mode) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1847) (ip->i_d.di_size != 0 || XFS_ISIZE(ip) != 0 ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1848) ip->i_df.if_nextents > 0 || ip->i_delayed_blks > 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1849) truncate = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1850)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1851) error = xfs_qm_dqattach(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1852) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1853) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1854)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1855) if (S_ISLNK(VFS_I(ip)->i_mode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1856) error = xfs_inactive_symlink(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1857) else if (truncate)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1858) error = xfs_inactive_truncate(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1859) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1860) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1861)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1862) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1863) * If there are attributes associated with the file then blow them away
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1864) * now. The code calls a routine that recursively deconstructs the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1865) * attribute fork. If also blows away the in-core attribute fork.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1866) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1867) if (XFS_IFORK_Q(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1868) error = xfs_attr_inactive(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1869) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1870) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1871) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1872)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1873) ASSERT(!ip->i_afp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1874) ASSERT(ip->i_d.di_forkoff == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1875)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1876) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1877) * Free the inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1878) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1879) error = xfs_inactive_ifree(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1880) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1881) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1882)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1883) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1884) * Release the dquots held by inode, if any.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1885) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1886) xfs_qm_dqdetach(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1887) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1888)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1889) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1890) * In-Core Unlinked List Lookups
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1891) * =============================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1892) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1893) * Every inode is supposed to be reachable from some other piece of metadata
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1894) * with the exception of the root directory. Inodes with a connection to a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1895) * file descriptor but not linked from anywhere in the on-disk directory tree
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1896) * are collectively known as unlinked inodes, though the filesystem itself
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1897) * maintains links to these inodes so that on-disk metadata are consistent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1898) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1899) * XFS implements a per-AG on-disk hash table of unlinked inodes. The AGI
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1900) * header contains a number of buckets that point to an inode, and each inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1901) * record has a pointer to the next inode in the hash chain. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1902) * singly-linked list causes scaling problems in the iunlink remove function
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1903) * because we must walk that list to find the inode that points to the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1904) * being removed from the unlinked hash bucket list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1905) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1906) * What if we modelled the unlinked list as a collection of records capturing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1907) * "X.next_unlinked = Y" relations? If we indexed those records on Y, we'd
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1908) * have a fast way to look up unlinked list predecessors, which avoids the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1909) * slow list walk. That's exactly what we do here (in-core) with a per-AG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1910) * rhashtable.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1911) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1912) * Because this is a backref cache, we ignore operational failures since the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1913) * iunlink code can fall back to the slow bucket walk. The only errors that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1914) * should bubble out are for obviously incorrect situations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1915) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1916) * All users of the backref cache MUST hold the AGI buffer lock to serialize
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1917) * access or have otherwise provided for concurrency control.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1918) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1919)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1920) /* Capture a "X.next_unlinked = Y" relationship. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1921) struct xfs_iunlink {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1922) struct rhash_head iu_rhash_head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1923) xfs_agino_t iu_agino; /* X */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1924) xfs_agino_t iu_next_unlinked; /* Y */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1925) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1926)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1927) /* Unlinked list predecessor lookup hashtable construction */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1928) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1929) xfs_iunlink_obj_cmpfn(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1930) struct rhashtable_compare_arg *arg,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1931) const void *obj)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1932) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1933) const xfs_agino_t *key = arg->key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1934) const struct xfs_iunlink *iu = obj;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1935)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1936) if (iu->iu_next_unlinked != *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1937) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1938) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1939) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1940)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1941) static const struct rhashtable_params xfs_iunlink_hash_params = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1942) .min_size = XFS_AGI_UNLINKED_BUCKETS,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1943) .key_len = sizeof(xfs_agino_t),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1944) .key_offset = offsetof(struct xfs_iunlink,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1945) iu_next_unlinked),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1946) .head_offset = offsetof(struct xfs_iunlink, iu_rhash_head),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1947) .automatic_shrinking = true,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1948) .obj_cmpfn = xfs_iunlink_obj_cmpfn,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1949) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1950)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1951) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1952) * Return X, where X.next_unlinked == @agino. Returns NULLAGINO if no such
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1953) * relation is found.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1954) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1955) static xfs_agino_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1956) xfs_iunlink_lookup_backref(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1957) struct xfs_perag *pag,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1958) xfs_agino_t agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1959) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1960) struct xfs_iunlink *iu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1961)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1962) iu = rhashtable_lookup_fast(&pag->pagi_unlinked_hash, &agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1963) xfs_iunlink_hash_params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1964) return iu ? iu->iu_agino : NULLAGINO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1965) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1966)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1967) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1968) * Take ownership of an iunlink cache entry and insert it into the hash table.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1969) * If successful, the entry will be owned by the cache; if not, it is freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1970) * Either way, the caller does not own @iu after this call.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1971) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1972) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1973) xfs_iunlink_insert_backref(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1974) struct xfs_perag *pag,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1975) struct xfs_iunlink *iu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1976) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1977) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1978)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1979) error = rhashtable_insert_fast(&pag->pagi_unlinked_hash,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1980) &iu->iu_rhash_head, xfs_iunlink_hash_params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1981) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1982) * Fail loudly if there already was an entry because that's a sign of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1983) * corruption of in-memory data. Also fail loudly if we see an error
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1984) * code we didn't anticipate from the rhashtable code. Currently we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1985) * only anticipate ENOMEM.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1986) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1987) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1988) WARN(error != -ENOMEM, "iunlink cache insert error %d", error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1989) kmem_free(iu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1990) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1991) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1992) * Absorb any runtime errors that aren't a result of corruption because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1993) * this is a cache and we can always fall back to bucket list scanning.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1994) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1995) if (error != 0 && error != -EEXIST)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1996) error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1997) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1998) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1999)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2000) /* Remember that @prev_agino.next_unlinked = @this_agino. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2001) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2002) xfs_iunlink_add_backref(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2003) struct xfs_perag *pag,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2004) xfs_agino_t prev_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2005) xfs_agino_t this_agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2006) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2007) struct xfs_iunlink *iu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2008)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2009) if (XFS_TEST_ERROR(false, pag->pag_mount, XFS_ERRTAG_IUNLINK_FALLBACK))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2010) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2011)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2012) iu = kmem_zalloc(sizeof(*iu), KM_NOFS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2013) iu->iu_agino = prev_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2014) iu->iu_next_unlinked = this_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2015)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2016) return xfs_iunlink_insert_backref(pag, iu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2017) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2018)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2019) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2020) * Replace X.next_unlinked = @agino with X.next_unlinked = @next_unlinked.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2021) * If @next_unlinked is NULLAGINO, we drop the backref and exit. If there
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2022) * wasn't any such entry then we don't bother.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2023) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2024) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2025) xfs_iunlink_change_backref(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2026) struct xfs_perag *pag,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2027) xfs_agino_t agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2028) xfs_agino_t next_unlinked)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2029) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2030) struct xfs_iunlink *iu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2031) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2032)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2033) /* Look up the old entry; if there wasn't one then exit. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2034) iu = rhashtable_lookup_fast(&pag->pagi_unlinked_hash, &agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2035) xfs_iunlink_hash_params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2036) if (!iu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2037) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2038)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2039) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2040) * Remove the entry. This shouldn't ever return an error, but if we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2041) * couldn't remove the old entry we don't want to add it again to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2042) * hash table, and if the entry disappeared on us then someone's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2043) * violated the locking rules and we need to fail loudly. Either way
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2044) * we cannot remove the inode because internal state is or would have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2045) * been corrupt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2046) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2047) error = rhashtable_remove_fast(&pag->pagi_unlinked_hash,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2048) &iu->iu_rhash_head, xfs_iunlink_hash_params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2049) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2050) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2051)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2052) /* If there is no new next entry just free our item and return. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2053) if (next_unlinked == NULLAGINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2054) kmem_free(iu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2055) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2056) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2057)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2058) /* Update the entry and re-add it to the hash table. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2059) iu->iu_next_unlinked = next_unlinked;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2060) return xfs_iunlink_insert_backref(pag, iu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2061) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2062)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2063) /* Set up the in-core predecessor structures. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2064) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2065) xfs_iunlink_init(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2066) struct xfs_perag *pag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2067) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2068) return rhashtable_init(&pag->pagi_unlinked_hash,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2069) &xfs_iunlink_hash_params);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2070) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2071)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2072) /* Free the in-core predecessor structures. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2073) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2074) xfs_iunlink_free_item(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2075) void *ptr,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2076) void *arg)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2077) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2078) struct xfs_iunlink *iu = ptr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2079) bool *freed_anything = arg;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2080)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2081) *freed_anything = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2082) kmem_free(iu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2083) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2084)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2085) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2086) xfs_iunlink_destroy(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2087) struct xfs_perag *pag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2088) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2089) bool freed_anything = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2090)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2091) rhashtable_free_and_destroy(&pag->pagi_unlinked_hash,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2092) xfs_iunlink_free_item, &freed_anything);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2093)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2094) ASSERT(freed_anything == false || XFS_FORCED_SHUTDOWN(pag->pag_mount));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2095) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2096)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2097) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2098) * Point the AGI unlinked bucket at an inode and log the results. The caller
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2099) * is responsible for validating the old value.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2100) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2101) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2102) xfs_iunlink_update_bucket(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2103) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2104) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2105) struct xfs_buf *agibp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2106) unsigned int bucket_index,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2107) xfs_agino_t new_agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2108) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2109) struct xfs_agi *agi = agibp->b_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2110) xfs_agino_t old_value;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2111) int offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2113) ASSERT(xfs_verify_agino_or_null(tp->t_mountp, agno, new_agino));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2114)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2115) old_value = be32_to_cpu(agi->agi_unlinked[bucket_index]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2116) trace_xfs_iunlink_update_bucket(tp->t_mountp, agno, bucket_index,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2117) old_value, new_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2118)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2119) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2120) * We should never find the head of the list already set to the value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2121) * passed in because either we're adding or removing ourselves from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2122) * head of the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2123) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2124) if (old_value == new_agino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2125) xfs_buf_mark_corrupt(agibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2126) return -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2127) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2128)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2129) agi->agi_unlinked[bucket_index] = cpu_to_be32(new_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2130) offset = offsetof(struct xfs_agi, agi_unlinked) +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2131) (sizeof(xfs_agino_t) * bucket_index);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2132) xfs_trans_log_buf(tp, agibp, offset, offset + sizeof(xfs_agino_t) - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2133) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2134) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2135)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2136) /* Set an on-disk inode's next_unlinked pointer. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2137) STATIC void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2138) xfs_iunlink_update_dinode(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2139) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2140) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2141) xfs_agino_t agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2142) struct xfs_buf *ibp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2143) struct xfs_dinode *dip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2144) struct xfs_imap *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2145) xfs_agino_t next_agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2146) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2147) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2148) int offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2149)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2150) ASSERT(xfs_verify_agino_or_null(mp, agno, next_agino));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2152) trace_xfs_iunlink_update_dinode(mp, agno, agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2153) be32_to_cpu(dip->di_next_unlinked), next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2154)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2155) dip->di_next_unlinked = cpu_to_be32(next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2156) offset = imap->im_boffset +
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2157) offsetof(struct xfs_dinode, di_next_unlinked);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2158)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2159) /* need to recalc the inode CRC if appropriate */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2160) xfs_dinode_calc_crc(mp, dip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2161) xfs_trans_inode_buf(tp, ibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2162) xfs_trans_log_buf(tp, ibp, offset, offset + sizeof(xfs_agino_t) - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2163) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2164)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2165) /* Set an in-core inode's unlinked pointer and return the old value. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2166) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2167) xfs_iunlink_update_inode(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2168) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2169) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2170) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2171) xfs_agino_t next_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2172) xfs_agino_t *old_next_agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2173) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2174) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2175) struct xfs_dinode *dip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2176) struct xfs_buf *ibp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2177) xfs_agino_t old_value;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2178) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2179)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2180) ASSERT(xfs_verify_agino_or_null(mp, agno, next_agino));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2182) error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2183) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2184) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2185)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2186) /* Make sure the old pointer isn't garbage. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2187) old_value = be32_to_cpu(dip->di_next_unlinked);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2188) if (!xfs_verify_agino_or_null(mp, agno, old_value)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2189) xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2190) sizeof(*dip), __this_address);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2191) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2192) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2193) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2195) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2196) * Since we're updating a linked list, we should never find that the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2197) * current pointer is the same as the new value, unless we're
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2198) * terminating the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2199) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2200) *old_next_agino = old_value;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2201) if (old_value == next_agino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2202) if (next_agino != NULLAGINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2203) xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2204) dip, sizeof(*dip), __this_address);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2205) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2206) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2207) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2208) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2210) /* Ok, update the new pointer. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2211) xfs_iunlink_update_dinode(tp, agno, XFS_INO_TO_AGINO(mp, ip->i_ino),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2212) ibp, dip, &ip->i_imap, next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2213) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2214) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2215) xfs_trans_brelse(tp, ibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2216) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2217) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2218)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2219) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2220) * This is called when the inode's link count has gone to 0 or we are creating
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2221) * a tmpfile via O_TMPFILE. The inode @ip must have nlink == 0.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2222) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2223) * We place the on-disk inode on a list in the AGI. It will be pulled from this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2224) * list when the inode is freed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2225) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2226) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2227) xfs_iunlink(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2228) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2229) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2230) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2231) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2232) struct xfs_agi *agi;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2233) struct xfs_buf *agibp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2234) xfs_agino_t next_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2235) xfs_agnumber_t agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2236) xfs_agino_t agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2237) short bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2238) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2240) ASSERT(VFS_I(ip)->i_nlink == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2241) ASSERT(VFS_I(ip)->i_mode != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2242) trace_xfs_iunlink(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2244) /* Get the agi buffer first. It ensures lock ordering on the list. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2245) error = xfs_read_agi(mp, tp, agno, &agibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2246) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2247) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2248) agi = agibp->b_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2250) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2251) * Get the index into the agi hash table for the list this inode will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2252) * go on. Make sure the pointer isn't garbage and that this inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2253) * isn't already on the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2254) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2255) next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2256) if (next_agino == agino ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2257) !xfs_verify_agino_or_null(mp, agno, next_agino)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2258) xfs_buf_mark_corrupt(agibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2259) return -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2260) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2261)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2262) if (next_agino != NULLAGINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2263) xfs_agino_t old_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2264)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2265) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2266) * There is already another inode in the bucket, so point this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2267) * inode to the current head of the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2268) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2269) error = xfs_iunlink_update_inode(tp, ip, agno, next_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2270) &old_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2271) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2272) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2273) ASSERT(old_agino == NULLAGINO);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2274)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2275) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2276) * agino has been unlinked, add a backref from the next inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2277) * back to agino.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2278) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2279) error = xfs_iunlink_add_backref(agibp->b_pag, agino, next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2280) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2281) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2282) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2283)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2284) /* Point the head of the list to point to this inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2285) return xfs_iunlink_update_bucket(tp, agno, agibp, bucket_index, agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2286) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2288) /* Return the imap, dinode pointer, and buffer for an inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2289) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2290) xfs_iunlink_map_ino(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2291) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2292) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2293) xfs_agino_t agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2294) struct xfs_imap *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2295) struct xfs_dinode **dipp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2296) struct xfs_buf **bpp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2297) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2298) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2299) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2300)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2301) imap->im_blkno = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2302) error = xfs_imap(mp, tp, XFS_AGINO_TO_INO(mp, agno, agino), imap, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2303) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2304) xfs_warn(mp, "%s: xfs_imap returned error %d.",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2305) __func__, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2306) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2307) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2308)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2309) error = xfs_imap_to_bp(mp, tp, imap, dipp, bpp, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2310) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2311) xfs_warn(mp, "%s: xfs_imap_to_bp returned error %d.",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2312) __func__, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2313) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2314) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2316) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2317) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2318)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2319) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2320) * Walk the unlinked chain from @head_agino until we find the inode that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2321) * points to @target_agino. Return the inode number, map, dinode pointer,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2322) * and inode cluster buffer of that inode as @agino, @imap, @dipp, and @bpp.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2323) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2324) * @tp, @pag, @head_agino, and @target_agino are input parameters.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2325) * @agino, @imap, @dipp, and @bpp are all output parameters.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2326) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2327) * Do not call this function if @target_agino is the head of the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2328) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2329) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2330) xfs_iunlink_map_prev(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2331) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2332) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2333) xfs_agino_t head_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2334) xfs_agino_t target_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2335) xfs_agino_t *agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2336) struct xfs_imap *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2337) struct xfs_dinode **dipp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2338) struct xfs_buf **bpp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2339) struct xfs_perag *pag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2340) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2341) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2342) xfs_agino_t next_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2343) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2344)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2345) ASSERT(head_agino != target_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2346) *bpp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2348) /* See if our backref cache can find it faster. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2349) *agino = xfs_iunlink_lookup_backref(pag, target_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2350) if (*agino != NULLAGINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2351) error = xfs_iunlink_map_ino(tp, agno, *agino, imap, dipp, bpp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2352) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2353) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2354)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2355) if (be32_to_cpu((*dipp)->di_next_unlinked) == target_agino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2356) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2357)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2358) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2359) * If we get here the cache contents were corrupt, so drop the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2360) * buffer and fall back to walking the bucket list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2361) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2362) xfs_trans_brelse(tp, *bpp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2363) *bpp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2364) WARN_ON_ONCE(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2365) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2367) trace_xfs_iunlink_map_prev_fallback(mp, agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2369) /* Otherwise, walk the entire bucket until we find it. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2370) next_agino = head_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2371) while (next_agino != target_agino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2372) xfs_agino_t unlinked_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2374) if (*bpp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2375) xfs_trans_brelse(tp, *bpp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2377) *agino = next_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2378) error = xfs_iunlink_map_ino(tp, agno, next_agino, imap, dipp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2379) bpp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2380) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2381) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2382)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2383) unlinked_agino = be32_to_cpu((*dipp)->di_next_unlinked);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2384) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2385) * Make sure this pointer is valid and isn't an obvious
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2386) * infinite loop.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2387) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2388) if (!xfs_verify_agino(mp, agno, unlinked_agino) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2389) next_agino == unlinked_agino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2390) XFS_CORRUPTION_ERROR(__func__,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2391) XFS_ERRLEVEL_LOW, mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2392) *dipp, sizeof(**dipp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2393) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2394) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2395) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2396) next_agino = unlinked_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2397) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2398)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2399) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2400) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2401)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2402) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2403) * Pull the on-disk inode from the AGI unlinked list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2404) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2405) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2406) xfs_iunlink_remove(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2407) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2408) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2409) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2410) struct xfs_mount *mp = tp->t_mountp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2411) struct xfs_agi *agi;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2412) struct xfs_buf *agibp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2413) struct xfs_buf *last_ibp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2414) struct xfs_dinode *last_dip = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2415) xfs_agnumber_t agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2416) xfs_agino_t agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2417) xfs_agino_t next_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2418) xfs_agino_t head_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2419) short bucket_index = agino % XFS_AGI_UNLINKED_BUCKETS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2420) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2422) trace_xfs_iunlink_remove(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2423)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2424) /* Get the agi buffer first. It ensures lock ordering on the list. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2425) error = xfs_read_agi(mp, tp, agno, &agibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2426) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2427) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2428) agi = agibp->b_addr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2430) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2431) * Get the index into the agi hash table for the list this inode will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2432) * go on. Make sure the head pointer isn't garbage.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2433) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2434) head_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2435) if (!xfs_verify_agino(mp, agno, head_agino)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2436) XFS_CORRUPTION_ERROR(__func__, XFS_ERRLEVEL_LOW, mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2437) agi, sizeof(*agi));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2438) return -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2439) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2441) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2442) * Set our inode's next_unlinked pointer to NULL and then return
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2443) * the old pointer value so that we can update whatever was previous
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2444) * to us in the list to point to whatever was next in the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2445) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2446) error = xfs_iunlink_update_inode(tp, ip, agno, NULLAGINO, &next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2447) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2448) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2449)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2450) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2451) * If there was a backref pointing from the next inode back to this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2452) * one, remove it because we've removed this inode from the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2453) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2454) * Later, if this inode was in the middle of the list we'll update
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2455) * this inode's backref to point from the next inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2456) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2457) if (next_agino != NULLAGINO) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2458) error = xfs_iunlink_change_backref(agibp->b_pag, next_agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2459) NULLAGINO);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2460) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2461) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2462) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2463)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2464) if (head_agino != agino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2465) struct xfs_imap imap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2466) xfs_agino_t prev_agino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2468) /* We need to search the list for the inode being freed. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2469) error = xfs_iunlink_map_prev(tp, agno, head_agino, agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2470) &prev_agino, &imap, &last_dip, &last_ibp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2471) agibp->b_pag);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2472) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2473) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2474)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2475) /* Point the previous inode on the list to the next inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2476) xfs_iunlink_update_dinode(tp, agno, prev_agino, last_ibp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2477) last_dip, &imap, next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2478)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2479) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2480) * Now we deal with the backref for this inode. If this inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2481) * pointed at a real inode, change the backref that pointed to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2482) * us to point to our old next. If this inode was the end of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2483) * the list, delete the backref that pointed to us. Note that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2484) * change_backref takes care of deleting the backref if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2485) * next_agino is NULLAGINO.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2486) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2487) return xfs_iunlink_change_backref(agibp->b_pag, agino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2488) next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2489) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2490)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2491) /* Point the head of the list to the next unlinked inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2492) return xfs_iunlink_update_bucket(tp, agno, agibp, bucket_index,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2493) next_agino);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2494) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2495)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2496) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2497) * Look up the inode number specified and if it is not already marked XFS_ISTALE
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2498) * mark it stale. We should only find clean inodes in this lookup that aren't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2499) * already stale.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2500) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2501) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2502) xfs_ifree_mark_inode_stale(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2503) struct xfs_buf *bp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2504) struct xfs_inode *free_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2505) xfs_ino_t inum)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2506) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2507) struct xfs_mount *mp = bp->b_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2508) struct xfs_perag *pag = bp->b_pag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2509) struct xfs_inode_log_item *iip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2510) struct xfs_inode *ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2511)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2512) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2513) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2514) ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, inum));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2515)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2516) /* Inode not in memory, nothing to do */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2517) if (!ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2518) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2519) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2520) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2521)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2522) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2523) * because this is an RCU protected lookup, we could find a recently
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2524) * freed or even reallocated inode during the lookup. We need to check
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2525) * under the i_flags_lock for a valid inode here. Skip it if it is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2526) * valid, the wrong inode or stale.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2527) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2528) spin_lock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2529) if (ip->i_ino != inum || __xfs_iflags_test(ip, XFS_ISTALE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2530) goto out_iflags_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2531)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2532) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2533) * Don't try to lock/unlock the current inode, but we _cannot_ skip the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2534) * other inodes that we did not find in the list attached to the buffer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2535) * and are not already marked stale. If we can't lock it, back off and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2536) * retry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2537) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2538) if (ip != free_ip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2539) if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2540) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2541) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2542) delay(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2543) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2544) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2545) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2546) ip->i_flags |= XFS_ISTALE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2547)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2548) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2549) * If the inode is flushing, it is already attached to the buffer. All
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2550) * we needed to do here is mark the inode stale so buffer IO completion
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2551) * will remove it from the AIL.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2552) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2553) iip = ip->i_itemp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2554) if (__xfs_iflags_test(ip, XFS_IFLUSHING)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2555) ASSERT(!list_empty(&iip->ili_item.li_bio_list));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2556) ASSERT(iip->ili_last_fields);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2557) goto out_iunlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2558) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2559)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2560) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2561) * Inodes not attached to the buffer can be released immediately.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2562) * Everything else has to go through xfs_iflush_abort() on journal
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2563) * commit as the flock synchronises removal of the inode from the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2564) * cluster buffer against inode reclaim.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2565) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2566) if (!iip || list_empty(&iip->ili_item.li_bio_list))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2567) goto out_iunlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2569) __xfs_iflags_set(ip, XFS_IFLUSHING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2570) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2571) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2573) /* we have a dirty inode in memory that has not yet been flushed. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2574) spin_lock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2575) iip->ili_last_fields = iip->ili_fields;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2576) iip->ili_fields = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2577) iip->ili_fsync_fields = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2578) spin_unlock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2579) ASSERT(iip->ili_last_fields);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2580)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2581) if (ip != free_ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2582) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2583) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2584)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2585) out_iunlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2586) if (ip != free_ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2587) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2588) out_iflags_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2589) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2590) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2591) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2592)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2593) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2594) * A big issue when freeing the inode cluster is that we _cannot_ skip any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2595) * inodes that are in memory - they all must be marked stale and attached to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2596) * the cluster buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2597) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2598) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2599) xfs_ifree_cluster(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2600) struct xfs_inode *free_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2601) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2602) struct xfs_icluster *xic)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2603) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2604) struct xfs_mount *mp = free_ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2605) struct xfs_ino_geometry *igeo = M_IGEO(mp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2606) struct xfs_buf *bp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2607) xfs_daddr_t blkno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2608) xfs_ino_t inum = xic->first_ino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2609) int nbufs;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2610) int i, j;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2611) int ioffset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2612) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2613)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2614) nbufs = igeo->ialloc_blks / igeo->blocks_per_cluster;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2615)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2616) for (j = 0; j < nbufs; j++, inum += igeo->inodes_per_cluster) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2617) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2618) * The allocation bitmap tells us which inodes of the chunk were
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2619) * physically allocated. Skip the cluster if an inode falls into
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2620) * a sparse region.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2621) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2622) ioffset = inum - xic->first_ino;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2623) if ((xic->alloc & XFS_INOBT_MASK(ioffset)) == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2624) ASSERT(ioffset % igeo->inodes_per_cluster == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2625) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2626) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2627)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2628) blkno = XFS_AGB_TO_DADDR(mp, XFS_INO_TO_AGNO(mp, inum),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2629) XFS_INO_TO_AGBNO(mp, inum));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2630)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2631) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2632) * We obtain and lock the backing buffer first in the process
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2633) * here to ensure dirty inodes attached to the buffer remain in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2634) * the flushing state while we mark them stale.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2635) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2636) * If we scan the in-memory inodes first, then buffer IO can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2637) * complete before we get a lock on it, and hence we may fail
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2638) * to mark all the active inodes on the buffer stale.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2639) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2640) error = xfs_trans_get_buf(tp, mp->m_ddev_targp, blkno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2641) mp->m_bsize * igeo->blocks_per_cluster,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2642) XBF_UNMAPPED, &bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2643) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2644) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2645)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2646) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2647) * This buffer may not have been correctly initialised as we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2648) * didn't read it from disk. That's not important because we are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2649) * only using to mark the buffer as stale in the log, and to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2650) * attach stale cached inodes on it. That means it will never be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2651) * dispatched for IO. If it is, we want to know about it, and we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2652) * want it to fail. We can acheive this by adding a write
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2653) * verifier to the buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2654) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2655) bp->b_ops = &xfs_inode_buf_ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2656)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2657) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2658) * Now we need to set all the cached clean inodes as XFS_ISTALE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2659) * too. This requires lookups, and will skip inodes that we've
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2660) * already marked XFS_ISTALE.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2661) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2662) for (i = 0; i < igeo->inodes_per_cluster; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2663) xfs_ifree_mark_inode_stale(bp, free_ip, inum + i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2664)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2665) xfs_trans_stale_inode_buf(tp, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2666) xfs_trans_binval(tp, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2667) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2668) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2669) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2670)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2671) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2672) * This is called to return an inode to the inode free list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2673) * The inode should already be truncated to 0 length and have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2674) * no pages associated with it. This routine also assumes that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2675) * the inode is already a part of the transaction.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2676) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2677) * The on-disk copy of the inode will have been added to the list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2678) * of unlinked inodes in the AGI. We need to remove the inode from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2679) * that list atomically with respect to freeing it here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2680) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2681) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2682) xfs_ifree(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2683) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2684) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2685) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2686) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2687) struct xfs_icluster xic = { 0 };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2688) struct xfs_inode_log_item *iip = ip->i_itemp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2689)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2690) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2691) ASSERT(VFS_I(ip)->i_nlink == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2692) ASSERT(ip->i_df.if_nextents == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2693) ASSERT(ip->i_d.di_size == 0 || !S_ISREG(VFS_I(ip)->i_mode));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2694) ASSERT(ip->i_d.di_nblocks == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2695)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2696) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2697) * Pull the on-disk inode from the AGI unlinked list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2698) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2699) error = xfs_iunlink_remove(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2700) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2701) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2702)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2703) error = xfs_difree(tp, ip->i_ino, &xic);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2704) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2705) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2706)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2707) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2708) * Free any local-format data sitting around before we reset the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2709) * data fork to extents format. Note that the attr fork data has
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2710) * already been freed by xfs_attr_inactive.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2711) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2712) if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2713) kmem_free(ip->i_df.if_u1.if_data);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2714) ip->i_df.if_u1.if_data = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2715) ip->i_df.if_bytes = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2716) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2717)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2718) VFS_I(ip)->i_mode = 0; /* mark incore inode as free */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2719) ip->i_d.di_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2720) ip->i_d.di_flags2 = ip->i_mount->m_ino_geo.new_diflags2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2721) ip->i_d.di_dmevmask = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2722) ip->i_d.di_forkoff = 0; /* mark the attr fork not in use */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2723) ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2724)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2725) /* Don't attempt to replay owner changes for a deleted inode */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2726) spin_lock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2727) iip->ili_fields &= ~(XFS_ILOG_AOWNER | XFS_ILOG_DOWNER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2728) spin_unlock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2729)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2730) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2731) * Bump the generation count so no one will be confused
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2732) * by reincarnations of this inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2733) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2734) VFS_I(ip)->i_generation++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2735) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2736)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2737) if (xic.deleted)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2738) error = xfs_ifree_cluster(ip, tp, &xic);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2739)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2740) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2741) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2742)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2743) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2744) * This is called to unpin an inode. The caller must have the inode locked
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2745) * in at least shared mode so that the buffer cannot be subsequently pinned
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2746) * once someone is waiting for it to be unpinned.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2747) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2748) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2749) xfs_iunpin(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2750) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2751) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2752) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2753)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2754) trace_xfs_inode_unpin_nowait(ip, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2755)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2756) /* Give the log a push to start the unpinning I/O */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2757) xfs_log_force_lsn(ip->i_mount, ip->i_itemp->ili_last_lsn, 0, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2758)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2759) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2760)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2761) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2762) __xfs_iunpin_wait(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2763) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2764) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2765) wait_queue_head_t *wq = bit_waitqueue(&ip->i_flags, __XFS_IPINNED_BIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2766) DEFINE_WAIT_BIT(wait, &ip->i_flags, __XFS_IPINNED_BIT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2767)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2768) xfs_iunpin(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2769)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2770) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2771) prepare_to_wait(wq, &wait.wq_entry, TASK_UNINTERRUPTIBLE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2772) if (xfs_ipincount(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2773) io_schedule();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2774) } while (xfs_ipincount(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2775) finish_wait(wq, &wait.wq_entry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2776) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2777)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2778) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2779) xfs_iunpin_wait(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2780) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2781) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2782) if (xfs_ipincount(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2783) __xfs_iunpin_wait(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2784) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2785)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2786) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2787) * Removing an inode from the namespace involves removing the directory entry
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2788) * and dropping the link count on the inode. Removing the directory entry can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2789) * result in locking an AGF (directory blocks were freed) and removing a link
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2790) * count can result in placing the inode on an unlinked list which results in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2791) * locking an AGI.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2792) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2793) * The big problem here is that we have an ordering constraint on AGF and AGI
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2794) * locking - inode allocation locks the AGI, then can allocate a new extent for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2795) * new inodes, locking the AGF after the AGI. Similarly, freeing the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2796) * removes the inode from the unlinked list, requiring that we lock the AGI
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2797) * first, and then freeing the inode can result in an inode chunk being freed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2798) * and hence freeing disk space requiring that we lock an AGF.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2799) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2800) * Hence the ordering that is imposed by other parts of the code is AGI before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2801) * AGF. This means we cannot remove the directory entry before we drop the inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2802) * reference count and put it on the unlinked list as this results in a lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2803) * order of AGF then AGI, and this can deadlock against inode allocation and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2804) * freeing. Therefore we must drop the link counts before we remove the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2805) * directory entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2806) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2807) * This is still safe from a transactional point of view - it is not until we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2808) * get to xfs_defer_finish() that we have the possibility of multiple
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2809) * transactions in this operation. Hence as long as we remove the directory
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2810) * entry and drop the link count in the first transaction of the remove
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2811) * operation, there are no transactional constraints on the ordering here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2812) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2813) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2814) xfs_remove(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2815) xfs_inode_t *dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2816) struct xfs_name *name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2817) xfs_inode_t *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2818) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2819) xfs_mount_t *mp = dp->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2820) xfs_trans_t *tp = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2821) int is_dir = S_ISDIR(VFS_I(ip)->i_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2822) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2823) uint resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2824)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2825) trace_xfs_remove(dp, name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2826)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2827) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2828) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2829)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2830) error = xfs_qm_dqattach(dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2831) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2832) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2833)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2834) error = xfs_qm_dqattach(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2835) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2836) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2837)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2838) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2839) * We try to get the real space reservation first,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2840) * allowing for directory btree deletion(s) implying
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2841) * possible bmap insert(s). If we can't get the space
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2842) * reservation then we use 0 instead, and avoid the bmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2843) * btree insert(s) in the directory code by, if the bmap
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2844) * insert tries to happen, instead trimming the LAST
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2845) * block from the directory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2846) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2847) resblks = XFS_REMOVE_SPACE_RES(mp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2848) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_remove, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2849) if (error == -ENOSPC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2850) resblks = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2851) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_remove, 0, 0, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2852) &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2853) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2854) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2855) ASSERT(error != -ENOSPC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2856) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2857) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2858)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2859) xfs_lock_two_inodes(dp, XFS_ILOCK_EXCL, ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2860)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2861) xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2862) xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2863)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2864) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2865) * If we're removing a directory perform some additional validation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2866) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2867) if (is_dir) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2868) ASSERT(VFS_I(ip)->i_nlink >= 2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2869) if (VFS_I(ip)->i_nlink != 2) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2870) error = -ENOTEMPTY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2871) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2872) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2873) if (!xfs_dir_isempty(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2874) error = -ENOTEMPTY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2875) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2876) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2877)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2878) /* Drop the link from ip's "..". */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2879) error = xfs_droplink(tp, dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2880) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2881) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2882)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2883) /* Drop the "." link from ip to self. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2884) error = xfs_droplink(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2885) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2886) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2887) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2888) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2889) * When removing a non-directory we need to log the parent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2890) * inode here. For a directory this is done implicitly
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2891) * by the xfs_droplink call for the ".." entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2892) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2893) xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2894) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2895) xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2896)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2897) /* Drop the link from dp to ip. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2898) error = xfs_droplink(tp, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2899) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2900) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2901)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2902) error = xfs_dir_removename(tp, dp, name, ip->i_ino, resblks);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2903) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2904) ASSERT(error != -ENOENT);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2905) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2906) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2907)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2908) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2909) * If this is a synchronous mount, make sure that the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2910) * remove transaction goes to disk before returning to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2911) * the user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2912) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2913) if (mp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2914) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2915)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2916) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2917) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2918) goto std_return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2919)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2920) if (is_dir && xfs_inode_is_filestream(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2921) xfs_filestream_deassociate(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2922)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2923) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2924)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2925) out_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2926) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2927) std_return:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2928) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2929) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2930)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2931) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2932) * Enter all inodes for a rename transaction into a sorted array.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2933) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2934) #define __XFS_SORT_INODES 5
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2935) STATIC void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2936) xfs_sort_for_rename(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2937) struct xfs_inode *dp1, /* in: old (source) directory inode */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2938) struct xfs_inode *dp2, /* in: new (target) directory inode */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2939) struct xfs_inode *ip1, /* in: inode of old entry */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2940) struct xfs_inode *ip2, /* in: inode of new entry */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2941) struct xfs_inode *wip, /* in: whiteout inode */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2942) struct xfs_inode **i_tab,/* out: sorted array of inodes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2943) int *num_inodes) /* in/out: inodes in array */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2944) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2945) int i, j;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2946)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2947) ASSERT(*num_inodes == __XFS_SORT_INODES);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2948) memset(i_tab, 0, *num_inodes * sizeof(struct xfs_inode *));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2949)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2950) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2951) * i_tab contains a list of pointers to inodes. We initialize
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2952) * the table here & we'll sort it. We will then use it to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2953) * order the acquisition of the inode locks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2954) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2955) * Note that the table may contain duplicates. e.g., dp1 == dp2.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2956) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2957) i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2958) i_tab[i++] = dp1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2959) i_tab[i++] = dp2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2960) i_tab[i++] = ip1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2961) if (ip2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2962) i_tab[i++] = ip2;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2963) if (wip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2964) i_tab[i++] = wip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2965) *num_inodes = i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2966)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2967) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2968) * Sort the elements via bubble sort. (Remember, there are at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2969) * most 5 elements to sort, so this is adequate.)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2970) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2971) for (i = 0; i < *num_inodes; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2972) for (j = 1; j < *num_inodes; j++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2973) if (i_tab[j]->i_ino < i_tab[j-1]->i_ino) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2974) struct xfs_inode *temp = i_tab[j];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2975) i_tab[j] = i_tab[j-1];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2976) i_tab[j-1] = temp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2977) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2978) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2979) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2980) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2981)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2982) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2983) xfs_finish_rename(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2984) struct xfs_trans *tp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2985) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2986) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2987) * If this is a synchronous mount, make sure that the rename transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2988) * goes to disk before returning to the user.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2989) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2990) if (tp->t_mountp->m_flags & (XFS_MOUNT_WSYNC|XFS_MOUNT_DIRSYNC))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2991) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2992)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2993) return xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2994) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2995)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2996) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2997) * xfs_cross_rename()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2998) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2999) * responsible for handling RENAME_EXCHANGE flag in renameat2() sytemcall
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3000) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3001) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3002) xfs_cross_rename(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3003) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3004) struct xfs_inode *dp1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3005) struct xfs_name *name1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3006) struct xfs_inode *ip1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3007) struct xfs_inode *dp2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3008) struct xfs_name *name2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3009) struct xfs_inode *ip2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3010) int spaceres)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3011) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3012) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3013) int ip1_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3014) int ip2_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3015) int dp2_flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3016)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3017) /* Swap inode number for dirent in first parent */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3018) error = xfs_dir_replace(tp, dp1, name1, ip2->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3019) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3020) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3021)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3022) /* Swap inode number for dirent in second parent */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3023) error = xfs_dir_replace(tp, dp2, name2, ip1->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3024) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3025) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3026)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3027) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3028) * If we're renaming one or more directories across different parents,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3029) * update the respective ".." entries (and link counts) to match the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3030) * parents.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3031) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3032) if (dp1 != dp2) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3033) dp2_flags = XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3034)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3035) if (S_ISDIR(VFS_I(ip2)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3036) error = xfs_dir_replace(tp, ip2, &xfs_name_dotdot,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3037) dp1->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3038) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3039) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3040)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3041) /* transfer ip2 ".." reference to dp1 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3042) if (!S_ISDIR(VFS_I(ip1)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3043) error = xfs_droplink(tp, dp2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3044) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3045) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3046) xfs_bumplink(tp, dp1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3047) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3048)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3049) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3050) * Although ip1 isn't changed here, userspace needs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3051) * to be warned about the change, so that applications
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3052) * relying on it (like backup ones), will properly
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3053) * notify the change
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3054) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3055) ip1_flags |= XFS_ICHGTIME_CHG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3056) ip2_flags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3057) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3058)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3059) if (S_ISDIR(VFS_I(ip1)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3060) error = xfs_dir_replace(tp, ip1, &xfs_name_dotdot,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3061) dp2->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3062) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3063) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3064)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3065) /* transfer ip1 ".." reference to dp2 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3066) if (!S_ISDIR(VFS_I(ip2)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3067) error = xfs_droplink(tp, dp1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3068) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3069) goto out_trans_abort;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3070) xfs_bumplink(tp, dp2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3071) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3072)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3073) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3074) * Although ip2 isn't changed here, userspace needs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3075) * to be warned about the change, so that applications
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3076) * relying on it (like backup ones), will properly
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3077) * notify the change
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3078) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3079) ip1_flags |= XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3080) ip2_flags |= XFS_ICHGTIME_CHG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3081) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3082) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3083)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3084) if (ip1_flags) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3085) xfs_trans_ichgtime(tp, ip1, ip1_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3086) xfs_trans_log_inode(tp, ip1, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3087) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3088) if (ip2_flags) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3089) xfs_trans_ichgtime(tp, ip2, ip2_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3090) xfs_trans_log_inode(tp, ip2, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3091) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3092) if (dp2_flags) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3093) xfs_trans_ichgtime(tp, dp2, dp2_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3094) xfs_trans_log_inode(tp, dp2, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3095) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3096) xfs_trans_ichgtime(tp, dp1, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3097) xfs_trans_log_inode(tp, dp1, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3098) return xfs_finish_rename(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3099)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3100) out_trans_abort:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3101) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3102) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3103) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3104)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3105) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3106) * xfs_rename_alloc_whiteout()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3107) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3108) * Return a referenced, unlinked, unlocked inode that can be used as a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3109) * whiteout in a rename transaction. We use a tmpfile inode here so that if we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3110) * crash between allocating the inode and linking it into the rename transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3111) * recovery will free the inode and we won't leak it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3112) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3113) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3114) xfs_rename_alloc_whiteout(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3115) struct xfs_inode *dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3116) struct xfs_inode **wip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3117) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3118) struct xfs_inode *tmpfile;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3119) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3120)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3121) error = xfs_create_tmpfile(dp, S_IFCHR | WHITEOUT_MODE, &tmpfile);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3122) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3123) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3124)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3125) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3126) * Prepare the tmpfile inode as if it were created through the VFS.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3127) * Complete the inode setup and flag it as linkable. nlink is already
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3128) * zero, so we can skip the drop_nlink.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3129) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3130) xfs_setup_iops(tmpfile);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3131) xfs_finish_inode_setup(tmpfile);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3132) VFS_I(tmpfile)->i_state |= I_LINKABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3133)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3134) *wip = tmpfile;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3135) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3136) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3138) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3139) * xfs_rename
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3140) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3141) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3142) xfs_rename(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3143) struct xfs_inode *src_dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3144) struct xfs_name *src_name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3145) struct xfs_inode *src_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3146) struct xfs_inode *target_dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3147) struct xfs_name *target_name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3148) struct xfs_inode *target_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3149) unsigned int flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3150) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3151) struct xfs_mount *mp = src_dp->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3152) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3153) struct xfs_inode *wip = NULL; /* whiteout inode */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3154) struct xfs_inode *inodes[__XFS_SORT_INODES];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3155) struct xfs_buf *agibp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3156) int num_inodes = __XFS_SORT_INODES;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3157) bool new_parent = (src_dp != target_dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3158) bool src_is_directory = S_ISDIR(VFS_I(src_ip)->i_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3159) int spaceres;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3160) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3161)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3162) trace_xfs_rename(src_dp, target_dp, src_name, target_name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3163)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3164) if ((flags & RENAME_EXCHANGE) && !target_ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3165) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3166)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3167) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3168) * If we are doing a whiteout operation, allocate the whiteout inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3169) * we will be placing at the target and ensure the type is set
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3170) * appropriately.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3171) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3172) if (flags & RENAME_WHITEOUT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3173) ASSERT(!(flags & (RENAME_NOREPLACE | RENAME_EXCHANGE)));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3174) error = xfs_rename_alloc_whiteout(target_dp, &wip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3175) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3176) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3178) /* setup target dirent info as whiteout */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3179) src_name->type = XFS_DIR3_FT_CHRDEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3180) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3182) xfs_sort_for_rename(src_dp, target_dp, src_ip, target_ip, wip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3183) inodes, &num_inodes);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3184)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3185) spaceres = XFS_RENAME_SPACE_RES(mp, target_name->len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3186) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, spaceres, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3187) if (error == -ENOSPC) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3188) spaceres = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3189) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_rename, 0, 0, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3190) &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3191) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3192) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3193) goto out_release_wip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3195) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3196) * Attach the dquots to the inodes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3197) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3198) error = xfs_qm_vop_rename_dqattach(inodes);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3199) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3200) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3201)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3202) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3203) * Lock all the participating inodes. Depending upon whether
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3204) * the target_name exists in the target directory, and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3205) * whether the target directory is the same as the source
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3206) * directory, we can lock from 2 to 4 inodes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3207) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3208) xfs_lock_inodes(inodes, num_inodes, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3210) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3211) * Join all the inodes to the transaction. From this point on,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3212) * we can rely on either trans_commit or trans_cancel to unlock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3213) * them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3214) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3215) xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3216) if (new_parent)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3217) xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3218) xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3219) if (target_ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3220) xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3221) if (wip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3222) xfs_trans_ijoin(tp, wip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3223)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3224) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3225) * If we are using project inheritance, we only allow renames
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3226) * into our tree when the project IDs are the same; else the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3227) * tree quota mechanism would be circumvented.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3228) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3229) if (unlikely((target_dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3230) target_dp->i_d.di_projid != src_ip->i_d.di_projid)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3231) error = -EXDEV;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3232) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3233) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3234)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3235) /* RENAME_EXCHANGE is unique from here on. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3236) if (flags & RENAME_EXCHANGE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3237) return xfs_cross_rename(tp, src_dp, src_name, src_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3238) target_dp, target_name, target_ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3239) spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3240)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3241) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3242) * Check for expected errors before we dirty the transaction
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3243) * so we can return an error without a transaction abort.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3244) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3245) if (target_ip == NULL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3246) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3247) * If there's no space reservation, check the entry will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3248) * fit before actually inserting it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3249) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3250) if (!spaceres) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3251) error = xfs_dir_canenter(tp, target_dp, target_name);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3252) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3253) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3254) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3255) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3256) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3257) * If target exists and it's a directory, check that whether
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3258) * it can be destroyed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3259) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3260) if (S_ISDIR(VFS_I(target_ip)->i_mode) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3261) (!xfs_dir_isempty(target_ip) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3262) (VFS_I(target_ip)->i_nlink > 2))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3263) error = -EEXIST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3264) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3265) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3266) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3268) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3269) * Directory entry creation below may acquire the AGF. Remove
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3270) * the whiteout from the unlinked list first to preserve correct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3271) * AGI/AGF locking order. This dirties the transaction so failures
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3272) * after this point will abort and log recovery will clean up the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3273) * mess.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3274) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3275) * For whiteouts, we need to bump the link count on the whiteout
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3276) * inode. After this point, we have a real link, clear the tmpfile
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3277) * state flag from the inode so it doesn't accidentally get misused
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3278) * in future.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3279) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3280) if (wip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3281) ASSERT(VFS_I(wip)->i_nlink == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3282) error = xfs_iunlink_remove(tp, wip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3283) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3284) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3286) xfs_bumplink(tp, wip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3287) VFS_I(wip)->i_state &= ~I_LINKABLE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3288) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3289)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3290) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3291) * Set up the target.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3292) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3293) if (target_ip == NULL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3294) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3295) * If target does not exist and the rename crosses
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3296) * directories, adjust the target directory link count
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3297) * to account for the ".." reference from the new entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3298) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3299) error = xfs_dir_createname(tp, target_dp, target_name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3300) src_ip->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3301) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3302) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3303)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3304) xfs_trans_ichgtime(tp, target_dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3305) XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3307) if (new_parent && src_is_directory) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3308) xfs_bumplink(tp, target_dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3309) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3310) } else { /* target_ip != NULL */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3311) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3312) * Link the source inode under the target name.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3313) * If the source inode is a directory and we are moving
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3314) * it across directories, its ".." entry will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3315) * inconsistent until we replace that down below.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3316) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3317) * In case there is already an entry with the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3318) * name at the destination directory, remove it first.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3319) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3321) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3322) * Check whether the replace operation will need to allocate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3323) * blocks. This happens when the shortform directory lacks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3324) * space and we have to convert it to a block format directory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3325) * When more blocks are necessary, we must lock the AGI first
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3326) * to preserve locking order (AGI -> AGF).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3327) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3328) if (xfs_dir2_sf_replace_needblock(target_dp, src_ip->i_ino)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3329) error = xfs_read_agi(mp, tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3330) XFS_INO_TO_AGNO(mp, target_ip->i_ino),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3331) &agibp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3332) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3333) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3334) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3335)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3336) error = xfs_dir_replace(tp, target_dp, target_name,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3337) src_ip->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3338) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3339) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3340)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3341) xfs_trans_ichgtime(tp, target_dp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3342) XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3344) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3345) * Decrement the link count on the target since the target
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3346) * dir no longer points to it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3347) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3348) error = xfs_droplink(tp, target_ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3349) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3350) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3351)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3352) if (src_is_directory) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3353) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3354) * Drop the link from the old "." entry.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3355) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3356) error = xfs_droplink(tp, target_ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3357) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3358) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3359) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3360) } /* target_ip != NULL */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3361)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3362) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3363) * Remove the source.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3364) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3365) if (new_parent && src_is_directory) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3366) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3367) * Rewrite the ".." entry to point to the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3368) * directory.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3369) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3370) error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3371) target_dp->i_ino, spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3372) ASSERT(error != -EEXIST);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3373) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3374) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3375) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3376)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3377) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3378) * We always want to hit the ctime on the source inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3379) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3380) * This isn't strictly required by the standards since the source
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3381) * inode isn't really being changed, but old unix file systems did
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3382) * it and some incremental backup programs won't work without it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3383) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3384) xfs_trans_ichgtime(tp, src_ip, XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3385) xfs_trans_log_inode(tp, src_ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3386)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3387) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3388) * Adjust the link count on src_dp. This is necessary when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3389) * renaming a directory, either within one parent when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3390) * the target existed, or across two parent directories.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3391) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3392) if (src_is_directory && (new_parent || target_ip != NULL)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3393)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3394) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3395) * Decrement link count on src_directory since the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3396) * entry that's moved no longer points to it.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3397) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3398) error = xfs_droplink(tp, src_dp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3399) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3400) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3401) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3402)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3403) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3404) * For whiteouts, we only need to update the source dirent with the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3405) * inode number of the whiteout inode rather than removing it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3406) * altogether.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3407) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3408) if (wip) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3409) error = xfs_dir_replace(tp, src_dp, src_name, wip->i_ino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3410) spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3411) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3412) error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3413) spaceres);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3414) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3415) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3416)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3417) xfs_trans_ichgtime(tp, src_dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3418) xfs_trans_log_inode(tp, src_dp, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3419) if (new_parent)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3420) xfs_trans_log_inode(tp, target_dp, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3422) error = xfs_finish_rename(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3423) if (wip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3424) xfs_irele(wip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3425) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3426)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3427) out_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3428) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3429) out_release_wip:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3430) if (wip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3431) xfs_irele(wip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3432) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3433) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3434)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3435) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3436) xfs_iflush(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3437) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3438) struct xfs_buf *bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3439) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3440) struct xfs_inode_log_item *iip = ip->i_itemp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3441) struct xfs_dinode *dip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3442) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3443) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3444)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3445) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3446) ASSERT(xfs_iflags_test(ip, XFS_IFLUSHING));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3447) ASSERT(ip->i_df.if_format != XFS_DINODE_FMT_BTREE ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3448) ip->i_df.if_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3449) ASSERT(iip->ili_item.li_buf == bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3450)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3451) dip = xfs_buf_offset(bp, ip->i_imap.im_boffset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3453) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3454) * We don't flush the inode if any of the following checks fail, but we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3455) * do still update the log item and attach to the backing buffer as if
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3456) * the flush happened. This is a formality to facilitate predictable
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3457) * error handling as the caller will shutdown and fail the buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3458) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3459) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3460) if (XFS_TEST_ERROR(dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3461) mp, XFS_ERRTAG_IFLUSH_1)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3462) xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3463) "%s: Bad inode %Lu magic number 0x%x, ptr "PTR_FMT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3464) __func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3465) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3466) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3467) if (S_ISREG(VFS_I(ip)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3468) if (XFS_TEST_ERROR(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3469) ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3470) ip->i_df.if_format != XFS_DINODE_FMT_BTREE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3471) mp, XFS_ERRTAG_IFLUSH_3)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3472) xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3473) "%s: Bad regular inode %Lu, ptr "PTR_FMT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3474) __func__, ip->i_ino, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3475) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3476) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3477) } else if (S_ISDIR(VFS_I(ip)->i_mode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3478) if (XFS_TEST_ERROR(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3479) ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3480) ip->i_df.if_format != XFS_DINODE_FMT_BTREE &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3481) ip->i_df.if_format != XFS_DINODE_FMT_LOCAL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3482) mp, XFS_ERRTAG_IFLUSH_4)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3483) xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3484) "%s: Bad directory inode %Lu, ptr "PTR_FMT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3485) __func__, ip->i_ino, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3486) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3487) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3488) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3489) if (XFS_TEST_ERROR(ip->i_df.if_nextents + xfs_ifork_nextents(ip->i_afp) >
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3490) ip->i_d.di_nblocks, mp, XFS_ERRTAG_IFLUSH_5)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3491) xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3492) "%s: detected corrupt incore inode %Lu, "
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3493) "total extents = %d, nblocks = %Ld, ptr "PTR_FMT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3494) __func__, ip->i_ino,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3495) ip->i_df.if_nextents + xfs_ifork_nextents(ip->i_afp),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3496) ip->i_d.di_nblocks, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3497) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3498) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3499) if (XFS_TEST_ERROR(ip->i_d.di_forkoff > mp->m_sb.sb_inodesize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3500) mp, XFS_ERRTAG_IFLUSH_6)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3501) xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3502) "%s: bad inode %Lu, forkoff 0x%x, ptr "PTR_FMT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3503) __func__, ip->i_ino, ip->i_d.di_forkoff, ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3504) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3505) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3506)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3507) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3508) * Inode item log recovery for v2 inodes are dependent on the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3509) * di_flushiter count for correct sequencing. We bump the flush
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3510) * iteration count so we can detect flushes which postdate a log record
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3511) * during recovery. This is redundant as we now log every change and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3512) * hence this can't happen but we need to still do it to ensure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3513) * backwards compatibility with old kernels that predate logging all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3514) * inode changes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3515) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3516) if (!xfs_sb_version_has_v3inode(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3517) ip->i_d.di_flushiter++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3518)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3519) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3520) * If there are inline format data / attr forks attached to this inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3521) * make sure they are not corrupt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3522) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3523) if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3524) xfs_ifork_verify_local_data(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3525) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3526) if (ip->i_afp && ip->i_afp->if_format == XFS_DINODE_FMT_LOCAL &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3527) xfs_ifork_verify_local_attr(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3528) goto flush_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3529)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3530) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3531) * Copy the dirty parts of the inode into the on-disk inode. We always
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3532) * copy out the core of the inode, because if the inode is dirty at all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3533) * the core must be.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3534) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3535) xfs_inode_to_disk(ip, dip, iip->ili_item.li_lsn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3536)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3537) /* Wrap, we never let the log put out DI_MAX_FLUSH */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3538) if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3539) ip->i_d.di_flushiter = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3540)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3541) xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3542) if (XFS_IFORK_Q(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3543) xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3544)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3545) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3546) * We've recorded everything logged in the inode, so we'd like to clear
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3547) * the ili_fields bits so we don't log and flush things unnecessarily.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3548) * However, we can't stop logging all this information until the data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3549) * we've copied into the disk buffer is written to disk. If we did we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3550) * might overwrite the copy of the inode in the log with all the data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3551) * after re-logging only part of it, and in the face of a crash we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3552) * wouldn't have all the data we need to recover.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3553) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3554) * What we do is move the bits to the ili_last_fields field. When
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3555) * logging the inode, these bits are moved back to the ili_fields field.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3556) * In the xfs_buf_inode_iodone() routine we clear ili_last_fields, since
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3557) * we know that the information those bits represent is permanently on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3558) * disk. As long as the flush completes before the inode is logged
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3559) * again, then both ili_fields and ili_last_fields will be cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3560) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3561) error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3562) flush_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3563) spin_lock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3564) iip->ili_last_fields = iip->ili_fields;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3565) iip->ili_fields = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3566) iip->ili_fsync_fields = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3567) spin_unlock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3568)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3569) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3570) * Store the current LSN of the inode so that we can tell whether the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3571) * item has moved in the AIL from xfs_buf_inode_iodone().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3572) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3573) xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3574) &iip->ili_item.li_lsn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3575)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3576) /* generate the checksum. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3577) xfs_dinode_calc_crc(mp, dip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3578) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3579) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3580)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3581) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3582) * Non-blocking flush of dirty inode metadata into the backing buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3583) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3584) * The caller must have a reference to the inode and hold the cluster buffer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3585) * locked. The function will walk across all the inodes on the cluster buffer it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3586) * can find and lock without blocking, and flush them to the cluster buffer.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3587) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3588) * On successful flushing of at least one inode, the caller must write out the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3589) * buffer and release it. If no inodes are flushed, -EAGAIN will be returned and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3590) * the caller needs to release the buffer. On failure, the filesystem will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3591) * shut down, the buffer will have been unlocked and released, and EFSCORRUPTED
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3592) * will be returned.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3593) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3594) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3595) xfs_iflush_cluster(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3596) struct xfs_buf *bp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3597) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3598) struct xfs_mount *mp = bp->b_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3599) struct xfs_log_item *lip, *n;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3600) struct xfs_inode *ip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3601) struct xfs_inode_log_item *iip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3602) int clcount = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3603) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3605) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3606) * We must use the safe variant here as on shutdown xfs_iflush_abort()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3607) * can remove itself from the list.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3608) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3609) list_for_each_entry_safe(lip, n, &bp->b_li_list, li_bio_list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3610) iip = (struct xfs_inode_log_item *)lip;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3611) ip = iip->ili_inode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3612)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3613) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3614) * Quick and dirty check to avoid locks if possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3615) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3616) if (__xfs_iflags_test(ip, XFS_IRECLAIM | XFS_IFLUSHING))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3617) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3618) if (xfs_ipincount(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3619) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3620)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3621) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3622) * The inode is still attached to the buffer, which means it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3623) * dirty but reclaim might try to grab it. Check carefully for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3624) * that, and grab the ilock while still holding the i_flags_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3625) * to guarantee reclaim will not be able to reclaim this inode
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3626) * once we drop the i_flags_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3627) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3628) spin_lock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3629) ASSERT(!__xfs_iflags_test(ip, XFS_ISTALE));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3630) if (__xfs_iflags_test(ip, XFS_IRECLAIM | XFS_IFLUSHING)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3631) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3632) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3633) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3634)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3635) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3636) * ILOCK will pin the inode against reclaim and prevent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3637) * concurrent transactions modifying the inode while we are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3638) * flushing the inode. If we get the lock, set the flushing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3639) * state before we drop the i_flags_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3640) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3641) if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3642) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3643) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3644) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3645) __xfs_iflags_set(ip, XFS_IFLUSHING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3646) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3647)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3648) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3649) * Abort flushing this inode if we are shut down because the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3650) * inode may not currently be in the AIL. This can occur when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3651) * log I/O failure unpins the inode without inserting into the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3652) * AIL, leaving a dirty/unpinned inode attached to the buffer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3653) * that otherwise looks like it should be flushed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3654) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3655) if (XFS_FORCED_SHUTDOWN(mp)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3656) xfs_iunpin_wait(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3657) xfs_iflush_abort(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3658) xfs_iunlock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3659) error = -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3660) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3661) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3662)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3663) /* don't block waiting on a log force to unpin dirty inodes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3664) if (xfs_ipincount(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3665) xfs_iflags_clear(ip, XFS_IFLUSHING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3666) xfs_iunlock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3667) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3668) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3669)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3670) if (!xfs_inode_clean(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3671) error = xfs_iflush(ip, bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3672) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3673) xfs_iflags_clear(ip, XFS_IFLUSHING);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3674) xfs_iunlock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3675) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3676) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3677) clcount++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3678) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3679)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3680) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3681) bp->b_flags |= XBF_ASYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3682) xfs_buf_ioend_fail(bp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3683) xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3684) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3685) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3686)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3687) if (!clcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3688) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3689)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3690) XFS_STATS_INC(mp, xs_icluster_flushcnt);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3691) XFS_STATS_ADD(mp, xs_icluster_flushinode, clcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3692) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3693)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3694) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3695)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3696) /* Release an inode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3697) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3698) xfs_irele(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3699) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3700) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3701) trace_xfs_irele(ip, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3702) iput(VFS_I(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3703) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3704)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3705) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3706) * Ensure all commited transactions touching the inode are written to the log.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3707) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3708) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3709) xfs_log_force_inode(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3710) struct xfs_inode *ip)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3711) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3712) xfs_lsn_t lsn = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3713)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3714) xfs_ilock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3715) if (xfs_ipincount(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3716) lsn = ip->i_itemp->ili_last_lsn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3717) xfs_iunlock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3718)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3719) if (!lsn)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3720) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3721) return xfs_log_force_lsn(ip->i_mount, lsn, XFS_LOG_SYNC, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3722) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3723)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3724) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3725) * Grab the exclusive iolock for a data copy from src to dest, making sure to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3726) * abide vfs locking order (lowest pointer value goes first) and breaking the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3727) * layout leases before proceeding. The loop is needed because we cannot call
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3728) * the blocking break_layout() with the iolocks held, and therefore have to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3729) * back out both locks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3730) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3731) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3732) xfs_iolock_two_inodes_and_break_layout(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3733) struct inode *src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3734) struct inode *dest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3735) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3736) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3737)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3738) if (src > dest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3739) swap(src, dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3740)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3741) retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3742) /* Wait to break both inodes' layouts before we start locking. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3743) error = break_layout(src, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3744) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3745) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3746) if (src != dest) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3747) error = break_layout(dest, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3748) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3749) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3750) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3751)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3752) /* Lock one inode and make sure nobody got in and leased it. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3753) inode_lock(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3754) error = break_layout(src, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3755) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3756) inode_unlock(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3757) if (error == -EWOULDBLOCK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3758) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3759) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3760) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3761)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3762) if (src == dest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3763) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3764)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3765) /* Lock the other inode and make sure nobody got in and leased it. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3766) inode_lock_nested(dest, I_MUTEX_NONDIR2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3767) error = break_layout(dest, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3768) if (error) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3769) inode_unlock(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3770) inode_unlock(dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3771) if (error == -EWOULDBLOCK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3772) goto retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3773) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3774) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3775)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3776) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3777) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3778)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3779) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3780) * Lock two inodes so that userspace cannot initiate I/O via file syscalls or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3781) * mmap activity.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3782) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3783) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3784) xfs_ilock2_io_mmap(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3785) struct xfs_inode *ip1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3786) struct xfs_inode *ip2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3787) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3788) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3789)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3790) ret = xfs_iolock_two_inodes_and_break_layout(VFS_I(ip1), VFS_I(ip2));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3791) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3792) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3793) if (ip1 == ip2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3794) xfs_ilock(ip1, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3795) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3796) xfs_lock_two_inodes(ip1, XFS_MMAPLOCK_EXCL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3797) ip2, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3798) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3799) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3800)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3801) /* Unlock both inodes to allow IO and mmap activity. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3802) void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3803) xfs_iunlock2_io_mmap(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3804) struct xfs_inode *ip1,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3805) struct xfs_inode *ip2)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3806) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3807) bool same_inode = (ip1 == ip2);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3808)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3809) xfs_iunlock(ip2, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3810) if (!same_inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3811) xfs_iunlock(ip1, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3812) inode_unlock(VFS_I(ip2));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3813) if (!same_inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3814) inode_unlock(VFS_I(ip1));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3815) }
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags