^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) 2016 Oracle. All Rights Reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * Author: Darrick J. Wong <darrick.wong@oracle.com>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include "xfs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) #include "xfs_fs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include "xfs_shared.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include "xfs_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) #include "xfs_log_format.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) #include "xfs_trans_resv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) #include "xfs_mount.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #include "xfs_defer.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include "xfs_inode.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include "xfs_trans.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include "xfs_bmap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include "xfs_bmap_util.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include "xfs_trace.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include "xfs_icache.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include "xfs_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include "xfs_refcount_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include "xfs_refcount.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include "xfs_bmap_btree.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include "xfs_trans_space.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include "xfs_bit.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include "xfs_alloc.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) #include "xfs_quota.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include "xfs_reflink.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include "xfs_iomap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include "xfs_sb.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include "xfs_ag_resv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) * Copy on Write of Shared Blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * XFS must preserve "the usual" file semantics even when two files share
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * the same physical blocks. This means that a write to one file must not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) * alter the blocks in a different file; the way that we'll do that is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) * through the use of a copy-on-write mechanism. At a high level, that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) * means that when we want to write to a shared block, we allocate a new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) * block, write the data to the new block, and if that succeeds we map the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) * new block into the file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) * XFS provides a "delayed allocation" mechanism that defers the allocation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) * of disk blocks to dirty-but-not-yet-mapped file blocks as long as
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) * possible. This reduces fragmentation by enabling the filesystem to ask
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) * for bigger chunks less often, which is exactly what we want for CoW.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) * The delalloc mechanism begins when the kernel wants to make a block
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) * writable (write_begin or page_mkwrite). If the offset is not mapped, we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) * create a delalloc mapping, which is a regular in-core extent, but without
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) * a real startblock. (For delalloc mappings, the startblock encodes both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) * a flag that this is a delalloc mapping, and a worst-case estimate of how
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) * many blocks might be required to put the mapping into the BMBT.) delalloc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) * mappings are a reservation against the free space in the filesystem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) * adjacent mappings can also be combined into fewer larger mappings.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) * As an optimization, the CoW extent size hint (cowextsz) creates
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) * outsized aligned delalloc reservations in the hope of landing out of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) * order nearby CoW writes in a single extent on disk, thereby reducing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) * fragmentation and improving future performance.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) * D: --RRRRRRSSSRRRRRRRR--- (data fork)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) * C: ------DDDDDDD--------- (CoW fork)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) * When dirty pages are being written out (typically in writepage), the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) * delalloc reservations are converted into unwritten mappings by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) * allocating blocks and replacing the delalloc mapping with real ones.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) * A delalloc mapping can be replaced by several unwritten ones if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) * free space is fragmented.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) * D: --RRRRRRSSSRRRRRRRR---
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) * C: ------UUUUUUU---------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) * We want to adapt the delalloc mechanism for copy-on-write, since the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) * write paths are similar. The first two steps (creating the reservation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) * and allocating the blocks) are exactly the same as delalloc except that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) * the mappings must be stored in a separate CoW fork because we do not want
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) * to disturb the mapping in the data fork until we're sure that the write
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) * succeeded. IO completion in this case is the process of removing the old
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) * mapping from the data fork and moving the new mapping from the CoW fork to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) * the data fork. This will be discussed shortly.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) * For now, unaligned directio writes will be bounced back to the page cache.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) * Block-aligned directio writes will use the same mechanism as buffered
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) * writes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) * Just prior to submitting the actual disk write requests, we convert
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) * the extents representing the range of the file actually being written
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) * (as opposed to extra pieces created for the cowextsize hint) to real
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) * extents. This will become important in the next step:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) * D: --RRRRRRSSSRRRRRRRR---
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) * C: ------UUrrUUU---------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) * CoW remapping must be done after the data block write completes,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) * because we don't want to destroy the old data fork map until we're sure
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) * the new block has been written. Since the new mappings are kept in a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) * separate fork, we can simply iterate these mappings to find the ones
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) * that cover the file blocks that we just CoW'd. For each extent, simply
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) * unmap the corresponding range in the data fork, map the new range into
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) * the data fork, and remove the extent from the CoW fork. Because of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) * the presence of the cowextsize hint, however, we must be careful
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) * only to remap the blocks that we've actually written out -- we must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) * never remap delalloc reservations nor CoW staging blocks that have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) * yet to be written. This corresponds exactly to the real extents in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) * the CoW fork:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) * D: --RRRRRRrrSRRRRRRRR---
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) * C: ------UU--UUU---------
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) * Since the remapping operation can be applied to an arbitrary file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) * range, we record the need for the remap step as a flag in the ioend
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) * instead of declaring a new IO type. This is required for direct io
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) * because we only have ioend for the whole dio, and we have to be able to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) * remember the presence of unwritten blocks and CoW blocks with a single
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) * ioend structure. Better yet, the more ground we can cover with one
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) * ioend, the better.
^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) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) * Given an AG extent, find the lowest-numbered run of shared blocks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) * within that range and return the range in fbno/flen. If
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) * find_end_of_shared is true, return the longest contiguous extent of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) * shared blocks. If there are no shared extents, fbno and flen will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) * be set to NULLAGBLOCK and 0, respectively.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) xfs_reflink_find_shared(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) struct xfs_mount *mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) xfs_agnumber_t agno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) xfs_agblock_t agbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) xfs_extlen_t aglen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) xfs_agblock_t *fbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) xfs_extlen_t *flen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) bool find_end_of_shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) struct xfs_buf *agbp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) struct xfs_btree_cur *cur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) error = xfs_alloc_read_agf(mp, tp, agno, 0, &agbp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) cur = xfs_refcountbt_init_cursor(mp, tp, agbp, agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) error = xfs_refcount_find_shared(cur, agbno, aglen, fbno, flen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) find_end_of_shared);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) xfs_btree_del_cursor(cur, error);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) xfs_trans_brelse(tp, agbp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) * Trim the mapping to the next block where there's a change in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * shared/unshared status. More specifically, this means that we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) * find the lowest-numbered extent of shared blocks that coincides with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) * the given block mapping. If the shared extent overlaps the start of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) * the mapping, trim the mapping to the end of the shared extent. If
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) * the shared region intersects the mapping, trim the mapping to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) * start of the shared extent. If there are no shared regions that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) * overlap, just return the original extent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) xfs_reflink_trim_around_shared(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) struct xfs_bmbt_irec *irec,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) bool *shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) xfs_agnumber_t agno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) xfs_agblock_t agbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) xfs_extlen_t aglen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) xfs_agblock_t fbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) xfs_extlen_t flen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) /* Holes, unwritten, and delalloc extents cannot be shared */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) if (!xfs_is_cow_inode(ip) || !xfs_bmap_is_written_extent(irec)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) *shared = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) trace_xfs_reflink_trim_around_shared(ip, irec);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) agno = XFS_FSB_TO_AGNO(ip->i_mount, irec->br_startblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) agbno = XFS_FSB_TO_AGBNO(ip->i_mount, irec->br_startblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) aglen = irec->br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) error = xfs_reflink_find_shared(ip->i_mount, NULL, agno, agbno,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) aglen, &fbno, &flen, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) *shared = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) if (fbno == NULLAGBLOCK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) /* No shared blocks at all. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) } else if (fbno == agbno) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) * The start of this extent is shared. Truncate the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) * mapping at the end of the shared region so that a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) * subsequent iteration starts at the start of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * unshared region.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) irec->br_blockcount = flen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) *shared = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) * There's a shared extent midway through this extent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) * Truncate the mapping at the start of the shared
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) * extent so that a subsequent iteration starts at the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) * start of the shared region.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) irec->br_blockcount = fbno - agbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) }
^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) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) xfs_bmap_trim_cow(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) struct xfs_bmbt_irec *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) bool *shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) /* We can't update any real extents in always COW mode. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) if (xfs_is_always_cow_inode(ip) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) !isnullstartblock(imap->br_startblock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) *shared = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) /* Trim the mapping to the nearest shared extent boundary. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) return xfs_reflink_trim_around_shared(ip, imap, shared);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) xfs_reflink_convert_cow_locked(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) xfs_fileoff_t offset_fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) xfs_filblks_t count_fsb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) struct xfs_iext_cursor icur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) struct xfs_bmbt_irec got;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) struct xfs_btree_cur *dummy_cur = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) int dummy_logflags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) if (!xfs_iext_lookup_extent(ip, ip->i_cowfp, offset_fsb, &icur, &got))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) if (got.br_startoff >= offset_fsb + count_fsb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) if (got.br_state == XFS_EXT_NORM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) if (WARN_ON_ONCE(isnullstartblock(got.br_startblock)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) xfs_trim_extent(&got, offset_fsb, count_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) if (!got.br_blockcount)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) got.br_state = XFS_EXT_NORM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) error = xfs_bmap_add_extent_unwritten_real(NULL, ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) XFS_COW_FORK, &icur, &dummy_cur, &got,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) &dummy_logflags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) } while (xfs_iext_next_extent(ip->i_cowfp, &icur, &got));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) /* Convert all of the unwritten CoW extents in a file's range to real ones. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) xfs_reflink_convert_cow(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) xfs_off_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) xfs_off_t count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) xfs_fileoff_t offset_fsb = XFS_B_TO_FSBT(mp, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) xfs_fileoff_t end_fsb = XFS_B_TO_FSB(mp, offset + count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) xfs_filblks_t count_fsb = end_fsb - offset_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) ASSERT(count != 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) error = xfs_reflink_convert_cow_locked(ip, offset_fsb, count_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) * Find the extent that maps the given range in the COW fork. Even if the extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) * is not shared we might have a preallocation for it in the COW fork. If so we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) * use it that rather than trigger a new allocation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) xfs_find_trim_cow_extent(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) struct xfs_bmbt_irec *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) struct xfs_bmbt_irec *cmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) bool *shared,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) bool *found)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) xfs_fileoff_t offset_fsb = imap->br_startoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) xfs_filblks_t count_fsb = imap->br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) struct xfs_iext_cursor icur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) *found = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) * If we don't find an overlapping extent, trim the range we need to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) * allocate to fit the hole we found.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) if (!xfs_iext_lookup_extent(ip, ip->i_cowfp, offset_fsb, &icur, cmap))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) cmap->br_startoff = offset_fsb + count_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) if (cmap->br_startoff > offset_fsb) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) xfs_trim_extent(imap, imap->br_startoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) cmap->br_startoff - imap->br_startoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) return xfs_bmap_trim_cow(ip, imap, shared);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) *shared = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) if (isnullstartblock(cmap->br_startblock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) xfs_trim_extent(imap, cmap->br_startoff, cmap->br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) /* real extent found - no need to allocate */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) xfs_trim_extent(cmap, offset_fsb, count_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) *found = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) /* Allocate all CoW reservations covering a range of blocks in a file. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) xfs_reflink_allocate_cow(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) struct xfs_bmbt_irec *imap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) struct xfs_bmbt_irec *cmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) bool *shared,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) uint *lockmode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) bool convert_now)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) xfs_fileoff_t offset_fsb = imap->br_startoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) xfs_filblks_t count_fsb = imap->br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) int nimaps, error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) bool found;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) xfs_filblks_t resaligned;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) xfs_extlen_t resblks = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) if (!ip->i_cowfp) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) ASSERT(!xfs_is_reflink_inode(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) xfs_ifork_init_cow(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) error = xfs_find_trim_cow_extent(ip, imap, cmap, shared, &found);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) if (error || !*shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) if (found)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) goto convert;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) resaligned = xfs_aligned_fsb_count(imap->br_startoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) imap->br_blockcount, xfs_get_cowextsz_hint(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) resblks = XFS_DIOSTRAT_SPACE_RES(mp, resaligned);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) xfs_iunlock(ip, *lockmode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) *lockmode = XFS_ILOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) xfs_ilock(ip, *lockmode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) error = xfs_qm_dqattach_locked(ip, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) * Check for an overlapping extent again now that we dropped the ilock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) error = xfs_find_trim_cow_extent(ip, imap, cmap, shared, &found);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) if (error || !*shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) if (found) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) goto convert;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) error = xfs_trans_reserve_quota_nblks(tp, ip, resblks, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) XFS_QMOPT_RES_REGBLKS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) goto out_trans_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) /* Allocate the entire reservation as unwritten blocks. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) nimaps = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) error = xfs_bmapi_write(tp, ip, imap->br_startoff, imap->br_blockcount,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) XFS_BMAPI_COWFORK | XFS_BMAPI_PREALLOC, 0, cmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) &nimaps);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) goto out_unreserve;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) xfs_inode_set_cowblocks_tag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) * Allocation succeeded but the requested range was not even partially
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) * satisfied? Bail out!
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) if (nimaps == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) return -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) convert:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) xfs_trim_extent(cmap, offset_fsb, count_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) * COW fork extents are supposed to remain unwritten until we're ready
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) * to initiate a disk write. For direct I/O we are going to write the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) * data and need the conversion, but for buffered writes we're done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) if (!convert_now || cmap->br_state == XFS_EXT_NORM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) trace_xfs_reflink_convert_cow(ip, cmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) return xfs_reflink_convert_cow_locked(ip, offset_fsb, count_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) out_unreserve:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) xfs_trans_unreserve_quota_nblks(tp, ip, (long)resblks, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441) XFS_QMOPT_RES_REGBLKS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) out_trans_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) * Cancel CoW reservations for some block range of an inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) * If cancel_real is true this function cancels all COW fork extents for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) * inode; if cancel_real is false, real extents are not cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) * Caller must have already joined the inode to the current transaction. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * inode will be joined to the transaction returned to the caller.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) xfs_reflink_cancel_cow_blocks(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) struct xfs_trans **tpp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) xfs_fileoff_t offset_fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) xfs_fileoff_t end_fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) bool cancel_real)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) struct xfs_bmbt_irec got, del;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) struct xfs_iext_cursor icur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) if (!xfs_inode_has_cow_data(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) if (!xfs_iext_lookup_extent_before(ip, ifp, &end_fsb, &icur, &got))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) /* Walk backwards until we're out of the I/O range... */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) while (got.br_startoff + got.br_blockcount > offset_fsb) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) del = got;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) xfs_trim_extent(&del, offset_fsb, end_fsb - offset_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) /* Extent delete may have bumped ext forward */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) if (!del.br_blockcount) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) xfs_iext_prev(ifp, &icur);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) goto next_extent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) trace_xfs_reflink_cancel_cow(ip, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) if (isnullstartblock(del.br_startblock)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) error = xfs_bmap_del_extent_delay(ip, XFS_COW_FORK,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) &icur, &got, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) } else if (del.br_state == XFS_EXT_UNWRITTEN || cancel_real) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) ASSERT((*tpp)->t_firstblock == NULLFSBLOCK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) /* Free the CoW orphan record. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) xfs_refcount_free_cow_extent(*tpp, del.br_startblock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) del.br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) xfs_bmap_add_free(*tpp, del.br_startblock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) del.br_blockcount, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) /* Roll the transaction */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) error = xfs_defer_finish(tpp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) /* Remove the mapping from the CoW fork. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) xfs_bmap_del_extent_cow(ip, &icur, &got, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) /* Remove the quota reservation */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) error = xfs_trans_reserve_quota_nblks(NULL, ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) -(long)del.br_blockcount, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) XFS_QMOPT_RES_REGBLKS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) /* Didn't do anything, push cursor back. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) xfs_iext_prev(ifp, &icur);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) next_extent:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) if (!xfs_iext_get_extent(ifp, &icur, &got))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) /* clear tag if cow fork is emptied */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) if (!ifp->if_bytes)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) xfs_inode_clear_cowblocks_tag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) * Cancel CoW reservations for some byte range of an inode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) * If cancel_real is true this function cancels all COW fork extents for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) * inode; if cancel_real is false, real extents are not cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) xfs_reflink_cancel_cow_range(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) xfs_off_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) xfs_off_t count,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) bool cancel_real)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) xfs_fileoff_t offset_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) xfs_fileoff_t end_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) trace_xfs_reflink_cancel_cow_range(ip, offset, count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) ASSERT(ip->i_cowfp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) if (count == NULLFILEOFF)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) end_fsb = NULLFILEOFF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) end_fsb = XFS_B_TO_FSB(ip->i_mount, offset + count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) /* Start a rolling transaction to remove the mappings */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_write,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) /* Scrape out the old CoW reservations */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) error = xfs_reflink_cancel_cow_blocks(ip, &tp, offset_fsb, end_fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) cancel_real);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) out_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) trace_xfs_reflink_cancel_cow_range_error(ip, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) * Remap part of the CoW fork into the data fork.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) * We aim to remap the range starting at @offset_fsb and ending at @end_fsb
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) * into the data fork; this function will remap what it can (at the end of the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) * range) and update @end_fsb appropriately. Each remap gets its own
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) * transaction because we can end up merging and splitting bmbt blocks for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) * every remap operation and we'd like to keep the block reservation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) * requirements as low as possible.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) xfs_reflink_end_cow_extent(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) xfs_fileoff_t offset_fsb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) xfs_fileoff_t *end_fsb)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) struct xfs_bmbt_irec got, del;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) struct xfs_iext_cursor icur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, XFS_COW_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) xfs_filblks_t rlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) unsigned int resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) /* No COW extents? That's easy! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) if (ifp->if_bytes == 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) *end_fsb = offset_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) XFS_TRANS_RESERVE, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) * Lock the inode. We have to ijoin without automatic unlock because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) * the lead transaction is the refcountbt record deletion; the data
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) * fork update follows as a deferred log item.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) * In case of racing, overlapping AIO writes no COW extents might be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) * left by the time I/O completes for the loser of the race. In that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) * case we are done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) if (!xfs_iext_lookup_extent_before(ip, ifp, end_fsb, &icur, &got) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) got.br_startoff + got.br_blockcount <= offset_fsb) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) *end_fsb = offset_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) * Structure copy @got into @del, then trim @del to the range that we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) * were asked to remap. We preserve @got for the eventual CoW fork
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) * deletion; from now on @del represents the mapping that we're
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) * actually remapping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) del = got;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) xfs_trim_extent(&del, offset_fsb, *end_fsb - offset_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) ASSERT(del.br_blockcount > 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) * Only remap real extents that contain data. With AIO, speculative
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) * preallocations can leak into the range we are called upon, and we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) * need to skip them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) if (!xfs_bmap_is_written_extent(&got)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) *end_fsb = del.br_startoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) goto out_cancel;
^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) /* Unmap the old blocks in the data fork. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) rlen = del.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) error = __xfs_bunmapi(tp, ip, del.br_startoff, &rlen, 0, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) /* Trim the extent to whatever got unmapped. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) xfs_trim_extent(&del, del.br_startoff + rlen, del.br_blockcount - rlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) trace_xfs_reflink_cow_remap(ip, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) /* Free the CoW orphan record. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) xfs_refcount_free_cow_extent(tp, del.br_startblock, del.br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) /* Map the new blocks into the data fork. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) xfs_bmap_map_extent(tp, ip, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) /* Charge this new data fork mapping to the on-disk quota. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_DELBCOUNT,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) (long)del.br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) /* Remove the mapping from the CoW fork. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) xfs_bmap_del_extent_cow(ip, &icur, &got, &del);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) /* Update the caller about how much progress we made. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) *end_fsb = del.br_startoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) out_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^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) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) * Remap parts of a file's data fork after a successful CoW.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) xfs_reflink_end_cow(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) xfs_off_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708) xfs_off_t count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) xfs_fileoff_t offset_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) xfs_fileoff_t end_fsb;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) trace_xfs_reflink_end_cow(ip, offset, count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) end_fsb = XFS_B_TO_FSB(ip->i_mount, offset + count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) * Walk backwards until we're out of the I/O range. The loop function
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) * repeatedly cycles the ILOCK to allocate one transaction per remapped
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) * extent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) * If we're being called by writeback then the pages will still
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) * have PageWriteback set, which prevents races with reflink remapping
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) * and truncate. Reflink remapping prevents races with writeback by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) * taking the iolock and mmaplock before flushing the pages and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) * remapping, which means there won't be any further writeback or page
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) * cache dirtying until the reflink completes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) * We should never have two threads issuing writeback for the same file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) * region. There are also have post-eof checks in the writeback
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733) * preparation code so that we don't bother writing out pages that are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) * about to be truncated.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) * If we're being called as part of directio write completion, the dio
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) * count is still elevated, which reflink and truncate will wait for.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738) * Reflink remapping takes the iolock and mmaplock and waits for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) * pending dio to finish, which should prevent any directio until the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) * remap completes. Multiple concurrent directio writes to the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) * region are handled by end_cow processing only occurring for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) * threads which succeed; the outcome of multiple overlapping direct
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) * writes is not well defined anyway.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) * It's possible that a buffered write and a direct write could collide
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) * here (the buffered write stumbles in after the dio flushes and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) * invalidates the page cache and immediately queues writeback), but we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) * have never supported this 100%. If either disk write succeeds the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) * blocks will be remapped.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) while (end_fsb > offset_fsb && !error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) error = xfs_reflink_end_cow_extent(ip, offset_fsb, &end_fsb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) trace_xfs_reflink_end_cow_error(ip, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) * Free leftover CoW reservations that didn't get cleaned out.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) xfs_reflink_recover_cow(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) struct xfs_mount *mp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) xfs_agnumber_t agno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) if (!xfs_sb_version_hasreflink(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) for (agno = 0; agno < mp->m_sb.sb_agcount; agno++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) error = xfs_refcount_recover_cow_leftovers(mp, agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) * Reflinking (Block) Ranges of Two Files Together
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) * First, ensure that the reflink flag is set on both inodes. The flag is an
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) * optimization to avoid unnecessary refcount btree lookups in the write path.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) * Now we can iteratively remap the range of extents (and holes) in src to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) * corresponding ranges in dest. Let drange and srange denote the ranges of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) * logical blocks in dest and src touched by the reflink operation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) * While the length of drange is greater than zero,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) * - Read src's bmbt at the start of srange ("imap")
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) * - If imap doesn't exist, make imap appear to start at the end of srange
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) * with zero length.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) * - If imap starts before srange, advance imap to start at srange.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) * - If imap goes beyond srange, truncate imap to end at the end of srange.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) * - Punch (imap start - srange start + imap len) blocks from dest at
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) * offset (drange start).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) * - If imap points to a real range of pblks,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) * > Increase the refcount of the imap's pblks
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) * > Map imap's pblks into dest at the offset
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) * (drange start + imap start - srange start)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) * - Advance drange and srange by (imap start - srange start + imap len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) * Finally, if the reflink made dest longer, update both the in-core and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) * on-disk file sizes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) * ASCII Art Demonstration:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) * Let's say we want to reflink this source file:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) * ----SSSSSSS-SSSSS----SSSSSS (src file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) * <-------------------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) * into this destination file:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) * --DDDDDDDDDDDDDDDDDDD--DDD (dest file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) * <-------------------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) * '-' means a hole, and 'S' and 'D' are written blocks in the src and dest.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) * Observe that the range has different logical offsets in either file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) * Consider that the first extent in the source file doesn't line up with our
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) * reflink range. Unmapping and remapping are separate operations, so we can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) * unmap more blocks from the destination file than we remap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) * ----SSSSSSS-SSSSS----SSSSSS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) * <------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) * --DDDDD---------DDDDD--DDD
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) * <------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) * Now remap the source extent into the destination file:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) * ----SSSSSSS-SSSSS----SSSSSS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) * <------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) * --DDDDD--SSSSSSSDDDDD--DDD
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) * <------->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) * Do likewise with the second hole and extent in our range. Holes in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) * unmap range don't affect our operation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) * ----SSSSSSS-SSSSS----SSSSSS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) * <---->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) * --DDDDD--SSSSSSS-SSSSS-DDD
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) * <---->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) * Finally, unmap and remap part of the third extent. This will increase the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) * size of the destination file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) * ----SSSSSSS-SSSSS----SSSSSS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) * <----->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) * --DDDDD--SSSSSSS-SSSSS----SSS
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) * <----->
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) * Once we update the destination file's i_size, we're done.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) * Ensure the reflink bit is set in both inodes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) xfs_reflink_set_inode_flag(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) struct xfs_inode *src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) struct xfs_inode *dest)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) struct xfs_mount *mp = src->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) if (xfs_is_reflink_inode(src) && xfs_is_reflink_inode(dest))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) goto out_error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) /* Lock both files against IO */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) if (src->i_ino == dest->i_ino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) xfs_ilock(src, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880) xfs_lock_two_inodes(src, XFS_ILOCK_EXCL, dest, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) if (!xfs_is_reflink_inode(src)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) trace_xfs_reflink_set_inode_flag(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) xfs_trans_ijoin(tp, src, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) src->i_d.di_flags2 |= XFS_DIFLAG2_REFLINK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) xfs_trans_log_inode(tp, src, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) xfs_ifork_init_cow(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) xfs_iunlock(src, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) if (src->i_ino == dest->i_ino)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) goto commit_flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) if (!xfs_is_reflink_inode(dest)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) trace_xfs_reflink_set_inode_flag(dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) xfs_trans_ijoin(tp, dest, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) dest->i_d.di_flags2 |= XFS_DIFLAG2_REFLINK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) xfs_trans_log_inode(tp, dest, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899) xfs_ifork_init_cow(dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901) xfs_iunlock(dest, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) commit_flags:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) goto out_error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) out_error:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) trace_xfs_reflink_set_inode_flag_error(dest, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 912) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 913)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 914) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) * Update destination inode size & cowextsize hint, if necessary.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) xfs_reflink_update_dest(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) struct xfs_inode *dest,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) xfs_off_t newlen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) xfs_extlen_t cowextsize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922) unsigned int remap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) struct xfs_mount *mp = dest->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) if (newlen <= i_size_read(VFS_I(dest)) && cowextsize == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_ichange, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) goto out_error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) xfs_ilock(dest, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) xfs_trans_ijoin(tp, dest, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) if (newlen > i_size_read(VFS_I(dest))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939) trace_xfs_reflink_update_inode_size(dest, newlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) i_size_write(VFS_I(dest), newlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) dest->i_d.di_size = newlen;
^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) if (cowextsize) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) dest->i_d.di_cowextsize = cowextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) dest->i_d.di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) xfs_trans_log_inode(tp, dest, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953) goto out_error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) out_error:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957) trace_xfs_reflink_update_inode_size_error(dest, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) * Do we have enough reserve in this AG to handle a reflink? The refcount
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) * btree already reserved all the space it needs, but the rmap btree can grow
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) * infinitely, so we won't allow more reflinks when the AG is down to the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) * btree reserves.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) xfs_reflink_ag_has_free_space(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) struct xfs_mount *mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) xfs_agnumber_t agno)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) struct xfs_perag *pag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975) if (!xfs_sb_version_hasrmapbt(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) pag = xfs_perag_get(mp, agno);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) if (xfs_ag_resv_critical(pag, XFS_AG_RESV_RMAPBT) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) xfs_ag_resv_critical(pag, XFS_AG_RESV_METADATA))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) error = -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) xfs_perag_put(pag);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983) return error;
^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) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) * Remap the given extent into the file. The dmap blockcount will be set to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) * the number of blocks that were actually remapped.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) xfs_reflink_remap_extent(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) struct xfs_bmbt_irec *dmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994) xfs_off_t new_isize)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) struct xfs_bmbt_irec smap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) xfs_off_t newlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) int64_t qres, qdelta;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001) unsigned int resblks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) bool smap_real;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) bool dmap_written = xfs_bmap_is_written_extent(dmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) int nimaps;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) /* Start a rolling transaction to switch the mappings */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) * Read what's currently mapped in the destination file into smap.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) * If smap isn't a hole, we will have to remove it before we can add
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) * dmap to the destination file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) nimaps = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) error = xfs_bmapi_read(ip, dmap->br_startoff, dmap->br_blockcount,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) &smap, &nimaps, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026) ASSERT(nimaps == 1 && smap.br_startoff == dmap->br_startoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) smap_real = xfs_bmap_is_real_extent(&smap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) * We can only remap as many blocks as the smaller of the two extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) * maps, because we can only remap one extent at a time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) dmap->br_blockcount = min(dmap->br_blockcount, smap.br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) ASSERT(dmap->br_blockcount == smap.br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) trace_xfs_reflink_remap_extent_dest(ip, &smap);
^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) * Two extents mapped to the same physical block must not have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040) * different states; that's filesystem corruption. Move on to the next
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) * extent if they're both holes or both the same physical extent.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) if (dmap->br_startblock == smap.br_startblock) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) if (dmap->br_state != smap.br_state)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049) /* If both extents are unwritten, leave them alone. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) if (dmap->br_state == XFS_EXT_UNWRITTEN &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) smap.br_state == XFS_EXT_UNWRITTEN)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054) /* No reflinking if the AG of the dest mapping is low on space. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055) if (dmap_written) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) error = xfs_reflink_ag_has_free_space(mp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) XFS_FSB_TO_AGNO(mp, dmap->br_startblock));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) * Compute quota reservation if we think the quota block counter for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) * this file could increase.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) * Adding a written extent to the extent map can cause a bmbt split,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067) * and removing a mapped extent from the extent can cause a bmbt split.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) * The two operations cannot both cause a split since they operate on
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) * the same index in the bmap btree, so we only need a reservation for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070) * one bmbt split if either thing is happening.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) * If we are mapping a written extent into the file, we need to have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073) * enough quota block count reservation to handle the blocks in that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) * extent. We log only the delta to the quota block counts, so if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) * extent we're unmapping also has blocks allocated to it, we don't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076) * need a quota reservation for the extent itself.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) * Note that if we're replacing a delalloc reservation with a written
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079) * extent, we have to take the full quota reservation because removing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) * the delalloc reservation gives the block count back to the quota
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) * count. This is suboptimal, but the VFS flushed the dest range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082) * before we started. That should have removed all the delalloc
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083) * reservations, but we code defensively.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) qres = qdelta = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) if (smap_real || dmap_written)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) qres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088) if (!smap_real && dmap_written)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) qres += dmap->br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090) if (qres > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) error = xfs_trans_reserve_quota_nblks(tp, ip, qres, 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) XFS_QMOPT_RES_REGBLKS);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1095) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1096)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1097) if (smap_real) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) * If the extent we're unmapping is backed by storage (written
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) * or not), unmap the extent and drop its refcount.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) xfs_bmap_unmap_extent(tp, ip, &smap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) xfs_refcount_decrease_extent(tp, &smap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) qdelta -= smap.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) } else if (smap.br_startblock == DELAYSTARTBLOCK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) xfs_filblks_t len = smap.br_blockcount;
^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) * If the extent we're unmapping is a delalloc reservation,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) * we can use the regular bunmapi function to release the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) * incore state. Dropping the delalloc reservation takes care
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112) * of the quota reservation for us.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) error = __xfs_bunmapi(NULL, ip, smap.br_startoff, &len, 0, 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) goto out_cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) ASSERT(len == 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121) * If the extent we're sharing is backed by written storage, increase
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) * its refcount and map it into the file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) if (dmap_written) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) xfs_refcount_increase_extent(tp, dmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) xfs_bmap_map_extent(tp, ip, dmap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) qdelta += dmap->br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, qdelta);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) /* Update dest isize if needed. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) newlen = XFS_FSB_TO_B(mp, dmap->br_startoff + dmap->br_blockcount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134) newlen = min_t(xfs_off_t, newlen, new_isize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) if (newlen > i_size_read(VFS_I(ip))) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) trace_xfs_reflink_update_inode_size(ip, newlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) i_size_write(VFS_I(ip), newlen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) ip->i_d.di_size = newlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) /* Commit everything and unlock. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) out_cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) trace_xfs_reflink_remap_extent_error(ip, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) /* Remap a range of one file to the other. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158) xfs_reflink_remap_blocks(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) struct xfs_inode *src,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) loff_t pos_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) struct xfs_inode *dest,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) loff_t pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) loff_t remap_len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) loff_t *remapped)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166) struct xfs_bmbt_irec imap;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) struct xfs_mount *mp = src->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168) xfs_fileoff_t srcoff = XFS_B_TO_FSBT(mp, pos_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) xfs_fileoff_t destoff = XFS_B_TO_FSBT(mp, pos_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) xfs_filblks_t len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171) xfs_filblks_t remapped_len = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) xfs_off_t new_isize = pos_out + remap_len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173) int nimaps;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) len = min_t(xfs_filblks_t, XFS_B_TO_FSB(mp, remap_len),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) XFS_MAX_FILEOFF);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) trace_xfs_reflink_remap_blocks(src, srcoff, len, dest, destoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181) while (len > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) unsigned int lock_mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) /* Read extent from the source file */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185) nimaps = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) lock_mode = xfs_ilock_data_map_shared(src);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) error = xfs_bmapi_read(src, srcoff, len, &imap, &nimaps, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) xfs_iunlock(src, lock_mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192) * The caller supposedly flushed all dirty pages in the source
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) * file range, which means that writeback should have allocated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) * or deleted all delalloc reservations in that range. If we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) * find one, that's a good sign that something is seriously
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) * wrong here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) ASSERT(nimaps == 1 && imap.br_startoff == srcoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) if (imap.br_startblock == DELAYSTARTBLOCK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200) ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) error = -EFSCORRUPTED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) trace_xfs_reflink_remap_extent_src(src, &imap);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207) /* Remap into the destination file at the given offset. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208) imap.br_startoff = destoff;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209) error = xfs_reflink_remap_extent(dest, &imap, new_isize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) if (fatal_signal_pending(current)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214) error = -EINTR;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218) /* Advance drange/srange */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219) srcoff += imap.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) destoff += imap.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) len -= imap.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222) remapped_len += imap.br_blockcount;
^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) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226) trace_xfs_reflink_remap_blocks_error(dest, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227) *remapped = min_t(loff_t, remap_len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) XFS_FSB_TO_B(src->i_mount, remapped_len));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) * If we're reflinking to a point past the destination file's EOF, we must
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234) * zero any speculative post-EOF preallocations that sit between the old EOF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235) * and the destination file offset.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238) xfs_reflink_zero_posteof(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240) loff_t pos)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) loff_t isize = i_size_read(VFS_I(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244) if (pos <= isize)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247) trace_xfs_zero_eof(ip, isize, pos - isize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) return iomap_zero_range(VFS_I(ip), isize, pos - isize, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249) &xfs_buffered_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) * Prepare two files for range cloning. Upon a successful return both inodes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254) * will have the iolock and mmaplock held, the page cache of the out file will
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) * be truncated, and any leases on the out file will have been broken. This
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) * function borrows heavily from xfs_file_aio_write_checks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258) * The VFS allows partial EOF blocks to "match" for dedupe even though it hasn't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259) * checked that the bytes beyond EOF physically match. Hence we cannot use the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) * EOF block in the source dedupe range because it's not a complete block match,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) * hence can introduce a corruption into the file that has it's block replaced.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263) * In similar fashion, the VFS file cloning also allows partial EOF blocks to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) * "block aligned" for the purposes of cloning entire files. However, if the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265) * source file range includes the EOF block and it lands within the existing EOF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) * of the destination file, then we can expose stale data from beyond the source
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267) * file EOF in the destination file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) * XFS doesn't support partial block sharing, so in both cases we have check
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270) * these cases ourselves. For dedupe, we can simply round the length to dedupe
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271) * down to the previous whole block and ignore the partial EOF block. While this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) * means we can't dedupe the last block of a file, this is an acceptible
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273) * tradeoff for simplicity on implementation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) * For cloning, we want to share the partial EOF block if it is also the new EOF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) * block of the destination file. If the partial EOF block lies inside the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277) * existing destination EOF, then we have to abort the clone to avoid exposing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278) * stale data in the destination file. Hence we reject these clone attempts with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279) * -EINVAL in this case.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) xfs_reflink_remap_prep(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283) struct file *file_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) loff_t pos_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285) struct file *file_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) loff_t pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) loff_t *len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288) unsigned int remap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) struct inode *inode_in = file_inode(file_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291) struct xfs_inode *src = XFS_I(inode_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) struct inode *inode_out = file_inode(file_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) struct xfs_inode *dest = XFS_I(inode_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296) /* Lock both files against IO */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297) ret = xfs_ilock2_io_mmap(src, dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301) /* Check file eligibility and prepare for block sharing. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302) ret = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) /* Don't reflink realtime inodes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304) if (XFS_IS_REALTIME_INODE(src) || XFS_IS_REALTIME_INODE(dest))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) /* Don't share DAX file data for now. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308) if (IS_DAX(inode_in) || IS_DAX(inode_out))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) ret = generic_remap_file_range_prep(file_in, pos_in, file_out, pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) len, remap_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313) if (ret || *len == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316) /* Attach dquots to dest inode before changing block map */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) ret = xfs_qm_dqattach(dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1320)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1321) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1322) * Zero existing post-eof speculative preallocations in the destination
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323) * file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325) ret = xfs_reflink_zero_posteof(dest, pos_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) /* Set flags and remap blocks. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330) ret = xfs_reflink_set_inode_flag(src, dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335) * If pos_out > EOF, we may have dirtied blocks between EOF and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336) * pos_out. In that case, we need to extend the flush and unmap to cover
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337) * from EOF to the end of the copy length.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) if (pos_out > XFS_ISIZE(dest)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) loff_t flen = *len + (pos_out - XFS_ISIZE(dest));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341) ret = xfs_flush_unmap_range(dest, XFS_ISIZE(dest), flen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343) ret = xfs_flush_unmap_range(dest, pos_out, *len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350) xfs_iunlock2_io_mmap(src, dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1352) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1353)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1354) /* Does this inode need the reflink flag? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356) xfs_reflink_inode_has_shared_extents(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) struct xfs_trans *tp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359) bool *has_shared)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361) struct xfs_bmbt_irec got;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) struct xfs_ifork *ifp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) xfs_agnumber_t agno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365) xfs_agblock_t agbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366) xfs_extlen_t aglen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) xfs_agblock_t rbno;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368) xfs_extlen_t rlen;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369) struct xfs_iext_cursor icur;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370) bool found;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373) ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) if (!(ifp->if_flags & XFS_IFEXTENTS)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) error = xfs_iread_extents(tp, ip, XFS_DATA_FORK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) *has_shared = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381) found = xfs_iext_lookup_extent(ip, ifp, 0, &icur, &got);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) while (found) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) if (isnullstartblock(got.br_startblock) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384) got.br_state != XFS_EXT_NORM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385) goto next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) agno = XFS_FSB_TO_AGNO(mp, got.br_startblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) agbno = XFS_FSB_TO_AGBNO(mp, got.br_startblock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388) aglen = got.br_blockcount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390) error = xfs_reflink_find_shared(mp, tp, agno, agbno, aglen,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) &rbno, &rlen, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) /* Is there still a shared block here? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395) if (rbno != NULLAGBLOCK) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) *has_shared = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) next:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1400) found = xfs_iext_next_extent(ifp, &icur, &got);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1401) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1402)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1403) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1404) }
^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) * Clear the inode reflink flag if there are no shared extents.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1408) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1409) * The caller is responsible for joining the inode to the transaction passed in.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1410) * The inode will be joined to the transaction that is returned to the caller.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1411) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1412) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1413) xfs_reflink_clear_inode_flag(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1414) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1415) struct xfs_trans **tpp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1416) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1417) bool needs_flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1418) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1419)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1420) ASSERT(xfs_is_reflink_inode(ip));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1421)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1422) error = xfs_reflink_inode_has_shared_extents(*tpp, ip, &needs_flag);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1423) if (error || needs_flag)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1424) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1425)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1426) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1427) * We didn't find any shared blocks so turn off the reflink flag.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1428) * First, get rid of any leftover CoW mappings.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1429) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1430) error = xfs_reflink_cancel_cow_blocks(ip, tpp, 0, XFS_MAX_FILEOFF,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1431) true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1432) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1433) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1434)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1435) /* Clear the inode flag. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1436) trace_xfs_reflink_unset_inode_flag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1437) ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1438) xfs_inode_clear_cowblocks_tag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1439) xfs_trans_log_inode(*tpp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1440)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1441) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1442) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1443)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1444) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1445) * Clear the inode reflink flag if there are no shared extents and the size
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1446) * hasn't changed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1447) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1448) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1449) xfs_reflink_try_clear_inode_flag(
^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_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1453) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1454) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1455)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1456) /* Start a rolling transaction to remove the mappings */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1457) error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1458) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1459) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1460)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1461) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1462) xfs_trans_ijoin(tp, ip, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1463)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1464) error = xfs_reflink_clear_inode_flag(ip, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1465) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1466) goto cancel;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1467)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1468) error = xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1469) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1470) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1471)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1472) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1473) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1474) cancel:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1475) xfs_trans_cancel(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1476) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1477) xfs_iunlock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1478) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1479) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1480)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1481) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1482) * Pre-COW all shared blocks within a given byte range of a file and turn off
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1483) * the reflink flag if we unshare all of the file's blocks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1484) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1485) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1486) xfs_reflink_unshare(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1487) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1488) xfs_off_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1489) xfs_off_t len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1490) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1491) struct inode *inode = VFS_I(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1492) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1493)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1494) if (!xfs_is_reflink_inode(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1495) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1496)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1497) trace_xfs_reflink_unshare(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1498)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1499) inode_dio_wait(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1500)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1501) error = iomap_file_unshare(inode, offset, len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1502) &xfs_buffered_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1503) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1504) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1505)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1506) error = filemap_write_and_wait_range(inode->i_mapping, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1507) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1508) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1509)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1510) /* Turn off the reflink flag if possible. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1511) error = xfs_reflink_try_clear_inode_flag(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1512) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1513) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1514) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1515)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1516) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1517) trace_xfs_reflink_unshare_error(ip, error, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1518) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1519) }
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags