^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * Copyright (c) 2000-2005 Silicon Graphics, Inc.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * All Rights Reserved.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) #include "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_inode.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #include "xfs_trans.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include "xfs_inode_item.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_dir2.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include "xfs_dir2_priv.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include "xfs_ioctl.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include "xfs_trace.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #include "xfs_log.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #include "xfs_icache.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #include "xfs_pnfs.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) #include "xfs_iomap.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) #include "xfs_reflink.h"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) #include <linux/falloc.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) #include <linux/backing-dev.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #include <linux/mman.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #include <linux/fadvise.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) static const struct vm_operations_struct xfs_file_vm_ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * Decide if the given file range is aligned to the size of the fundamental
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * allocation unit for the file.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) static bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) xfs_is_falloc_aligned(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) loff_t pos,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) long long int len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) uint64_t mask;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) if (XFS_IS_REALTIME_INODE(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) if (!is_power_of_2(mp->m_sb.sb_rextsize)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) u64 rextbytes;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) u32 mod;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) rextbytes = XFS_FSB_TO_B(mp, mp->m_sb.sb_rextsize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) div_u64_rem(pos, rextbytes, &mod);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) if (mod)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) div_u64_rem(len, rextbytes, &mod);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) return mod == 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) mask = XFS_FSB_TO_B(mp, mp->m_sb.sb_rextsize) - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) mask = mp->m_sb.sb_blocksize - 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) return !((pos | len) & mask);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) xfs_update_prealloc_flags(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) struct xfs_inode *ip,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) enum xfs_prealloc_flags flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) struct xfs_trans *tp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_writeid,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) 0, 0, 0, &tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) xfs_ilock(ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) if (!(flags & XFS_PREALLOC_INVISIBLE)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) VFS_I(ip)->i_mode &= ~S_ISUID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) if (VFS_I(ip)->i_mode & S_IXGRP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) VFS_I(ip)->i_mode &= ~S_ISGID;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) if (flags & XFS_PREALLOC_SET)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) ip->i_d.di_flags |= XFS_DIFLAG_PREALLOC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) if (flags & XFS_PREALLOC_CLEAR)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) ip->i_d.di_flags &= ~XFS_DIFLAG_PREALLOC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) if (flags & XFS_PREALLOC_SYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) xfs_trans_set_sync(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99) return xfs_trans_commit(tp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) * Fsync operations on directories are much simpler than on regular files,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) * as there is no file data to flush, and thus also no need for explicit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) * cache flush operations, and there are no non-transaction metadata updates
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) * on directories either.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) xfs_dir_fsync(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) loff_t start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) loff_t end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) int datasync)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) struct xfs_inode *ip = XFS_I(file->f_mapping->host);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) trace_xfs_dir_fsync(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) return xfs_log_force_inode(ip);
^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) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) xfs_file_fsync(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) loff_t start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) loff_t end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) int datasync)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) struct inode *inode = file->f_mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) struct xfs_inode_log_item *iip = ip->i_itemp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) int error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) int log_flushed = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) xfs_lsn_t lsn = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) trace_xfs_file_fsync(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) error = file_write_and_wait_range(file, start, end);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) xfs_iflags_clear(ip, XFS_ITRUNCATED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) * If we have an RT and/or log subvolume we need to make sure to flush
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) * the write cache the device used for file data first. This is to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) * ensure newly written file data make it to disk before logging the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) * inode size in case of an extending write.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) if (XFS_IS_REALTIME_INODE(ip))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) xfs_blkdev_issue_flush(mp->m_rtdev_targp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) else if (mp->m_logdev_targp != mp->m_ddev_targp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) xfs_blkdev_issue_flush(mp->m_ddev_targp);
^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) * All metadata updates are logged, which means that we just have to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) * flush the log up to the latest LSN that touched the inode. If we have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) * concurrent fsync/fdatasync() calls, we need them to all block on the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) * log force before we clear the ili_fsync_fields field. This ensures
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) * that we don't get a racing sync operation that does not wait for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) * metadata to hit the journal before returning. If we race with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) * clearing the ili_fsync_fields, then all that will happen is the log
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) * force will do nothing as the lsn will already be on disk. We can't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) * race with setting ili_fsync_fields because that is done under
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) * XFS_ILOCK_EXCL, and that can't happen because we hold the lock shared
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) * until after the ili_fsync_fields is cleared.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) xfs_ilock(ip, XFS_ILOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) if (xfs_ipincount(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) if (!datasync ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) (iip->ili_fsync_fields & ~XFS_ILOG_TIMESTAMP))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) lsn = iip->ili_last_lsn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) if (lsn) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) error = xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, &log_flushed);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) spin_lock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) iip->ili_fsync_fields = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) spin_unlock(&iip->ili_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) xfs_iunlock(ip, XFS_ILOCK_SHARED);
^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) * If we only have a single device, and the log force about was
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) * a no-op we might have to flush the data device cache here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) * This can only happen for fdatasync/O_DSYNC if we were overwriting
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) * an already allocated file and thus do not have any metadata to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) * commit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) if (!log_flushed && !XFS_IS_REALTIME_INODE(ip) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) mp->m_logdev_targp == mp->m_ddev_targp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) xfs_blkdev_issue_flush(mp->m_ddev_targp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) xfs_file_dio_aio_read(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) struct iov_iter *to)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) struct xfs_inode *ip = XFS_I(file_inode(iocb->ki_filp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) size_t count = iov_iter_count(to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) ssize_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) trace_xfs_file_direct_read(ip, count, iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) if (!count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) return 0; /* skip atime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) file_accessed(iocb->ki_filp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) if (iocb->ki_flags & IOCB_NOWAIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) xfs_ilock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) ret = iomap_dio_rw(iocb, to, &xfs_read_iomap_ops, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) is_sync_kiocb(iocb));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) xfs_iunlock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) static noinline ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) xfs_file_dax_read(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) struct iov_iter *to)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) struct xfs_inode *ip = XFS_I(iocb->ki_filp->f_mapping->host);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) size_t count = iov_iter_count(to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) ssize_t ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) trace_xfs_file_dax_read(ip, count, iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) if (!count)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) return 0; /* skip atime */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) if (iocb->ki_flags & IOCB_NOWAIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) xfs_ilock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) ret = dax_iomap_rw(iocb, to, &xfs_read_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) xfs_iunlock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) file_accessed(iocb->ki_filp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) xfs_file_buffered_aio_read(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) struct iov_iter *to)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) struct xfs_inode *ip = XFS_I(file_inode(iocb->ki_filp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) ssize_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) trace_xfs_file_buffered_read(ip, iov_iter_count(to), iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) if (iocb->ki_flags & IOCB_NOWAIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) if (!xfs_ilock_nowait(ip, XFS_IOLOCK_SHARED))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) xfs_ilock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) ret = generic_file_read_iter(iocb, to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) xfs_iunlock(ip, XFS_IOLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) return ret;
^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) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) xfs_file_read_iter(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) struct iov_iter *to)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) struct inode *inode = file_inode(iocb->ki_filp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) struct xfs_mount *mp = XFS_I(inode)->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) ssize_t ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) XFS_STATS_INC(mp, xs_read_calls);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) if (IS_DAX(inode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) ret = xfs_file_dax_read(iocb, to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) else if (iocb->ki_flags & IOCB_DIRECT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) ret = xfs_file_dio_aio_read(iocb, to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) ret = xfs_file_buffered_aio_read(iocb, to);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) if (ret > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) XFS_STATS_ADD(mp, xs_read_bytes, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) * Common pre-write limit and setup checks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) * Called with the iolocked held either shared and exclusive according to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) * @iolock, and returns with it held. Might upgrade the iolock to exclusive
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) * if called for a direct write beyond i_size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) xfs_file_aio_write_checks(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) struct iov_iter *from,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) int *iolock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) struct file *file = iocb->ki_filp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) struct inode *inode = file->f_mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) ssize_t error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) size_t count = iov_iter_count(from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) bool drained_dio = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) loff_t isize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) restart:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) error = generic_write_checks(iocb, from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) if (error <= 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) error = xfs_break_layouts(inode, iolock, BREAK_WRITE);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) * For changing security info in file_remove_privs() we need i_rwsem
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) * exclusively.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) if (*iolock == XFS_IOLOCK_SHARED && !IS_NOSEC(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) xfs_iunlock(ip, *iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) *iolock = XFS_IOLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) xfs_ilock(ip, *iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) goto restart;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) * If the offset is beyond the size of the file, we need to zero any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) * blocks that fall between the existing EOF and the start of this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) * write. If zeroing is needed and we are currently holding the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) * iolock shared, we need to update it to exclusive which implies
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) * having to redo all checks before.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) * We need to serialise against EOF updates that occur in IO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) * completions here. We want to make sure that nobody is changing the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) * size while we do this check until we have placed an IO barrier (i.e.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) * hold the XFS_IOLOCK_EXCL) that prevents new IO from being dispatched.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) * The spinlock effectively forms a memory barrier once we have the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) * XFS_IOLOCK_EXCL so we are guaranteed to see the latest EOF value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) * and hence be able to correctly determine if we need to run zeroing.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) spin_lock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) isize = i_size_read(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) if (iocb->ki_pos > isize) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) if (!drained_dio) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) if (*iolock == XFS_IOLOCK_SHARED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) xfs_iunlock(ip, *iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) *iolock = XFS_IOLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368) xfs_ilock(ip, *iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) iov_iter_reexpand(from, count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) * We now have an IO submission barrier in place, but
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) * AIO can do EOF updates during IO completion and hence
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) * we now need to wait for all of them to drain. Non-AIO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) * DIO will have drained before we are given the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) * XFS_IOLOCK_EXCL, and so for most cases this wait is a
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) * no-op.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379) inode_dio_wait(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) drained_dio = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) goto restart;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) trace_xfs_zero_eof(ip, isize, iocb->ki_pos - isize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385) error = iomap_zero_range(inode, isize, iocb->ki_pos - isize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) NULL, &xfs_buffered_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) } else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) * Updating the timestamps will grab the ilock again from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) * xfs_fs_dirty_inode, so we have to call it after dropping the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) * lock above. Eventually we should look into a way to avoid
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) * the pointless lock roundtrip.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) return file_modified(file);
^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) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) xfs_dio_write_end_io(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404) ssize_t size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) int error,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) unsigned flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) struct inode *inode = file_inode(iocb->ki_filp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) loff_t offset = iocb->ki_pos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) unsigned int nofs_flag;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413) trace_xfs_end_io_direct_write(ip, offset, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) if (XFS_FORCED_SHUTDOWN(ip->i_mount))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417)
^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) if (!size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) * Capture amount written on completion as we can't reliably account
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) * for it on submission.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) XFS_STATS_ADD(ip->i_mount, xs_write_bytes, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) * We can allocate memory here while doing writeback on behalf of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) * memory reclaim. To avoid memory allocation deadlocks set the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432) * task-wide nofs context for the following operations.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) nofs_flag = memalloc_nofs_save();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) if (flags & IOMAP_DIO_COW) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) error = xfs_reflink_end_cow(ip, offset, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) * Unwritten conversion updates the in-core isize after extent
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) * conversion but before updating the on-disk size. Updating isize any
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445) * earlier allows a racing dio read to find unwritten extents before
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) * they are converted.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) if (flags & IOMAP_DIO_UNWRITTEN) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) error = xfs_iomap_write_unwritten(ip, offset, size, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454) * We need to update the in-core inode size here so that we don't end up
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) * with the on-disk inode size being outside the in-core inode size. We
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) * have no other method of updating EOF for AIO, so always do it here
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) * if necessary.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) * We need to lock the test/set EOF update as we can be racing with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) * other IO completions here to update the EOF. Failing to serialise
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) * here can result in EOF moving backwards and Bad Things Happen when
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462) * that occurs.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) spin_lock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) if (offset + size > i_size_read(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466) i_size_write(inode, offset + size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) error = xfs_setfilesize(ip, offset, size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) spin_unlock(&ip->i_flags_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) memalloc_nofs_restore(nofs_flag);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) static const struct iomap_dio_ops xfs_dio_write_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) .end_io = xfs_dio_write_end_io,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) * xfs_file_dio_aio_write - handle direct IO writes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) * Lock the inode appropriately to prepare for and issue a direct IO write.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) * By separating it from the buffered write path we remove all the tricky to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) * follow locking changes and looping.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) * If there are cached pages or we're extending the file, we need IOLOCK_EXCL
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490) * until we're sure the bytes at the new EOF have been zeroed and/or the cached
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) * pages are flushed out.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) * In most cases the direct IO writes will be done holding IOLOCK_SHARED
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494) * allowing them to be done in parallel with reads and other direct IO writes.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) * However, if the IO is not aligned to filesystem blocks, the direct IO layer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) * needs to do sub-block zeroing and that requires serialisation against other
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) * direct IOs to the same block. In this case we need to serialise the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) * submission of the unaligned IOs so that we don't get racing block zeroing in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) * the dio layer. To avoid the problem with aio, we also need to wait for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500) * outstanding IOs to complete so that unwritten extent conversion is completed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) * before we try to map the overlapping block. This is currently implemented by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) * hitting it with a big hammer (i.e. inode_dio_wait()).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) * Returns with locks held indicated by @iolock and errors indicated by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505) * negative return values.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) xfs_file_dio_aio_write(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) struct iov_iter *from)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) struct file *file = iocb->ki_filp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513) struct address_space *mapping = file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) struct inode *inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) struct xfs_mount *mp = ip->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) ssize_t ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518) int unaligned_io = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) int iolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) size_t count = iov_iter_count(from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) struct xfs_buftarg *target = xfs_inode_buftarg(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) /* DIO must be aligned to device logical sector size */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524) if ((iocb->ki_pos | count) & target->bt_logical_sectormask)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) * Don't take the exclusive iolock here unless the I/O is unaligned to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) * the file system block size. We don't need to consider the EOF
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) * extension case here because xfs_file_aio_write_checks() will relock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531) * the inode as necessary for EOF zeroing cases and fill out the new
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) * inode size as appropriate.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534) if ((iocb->ki_pos & mp->m_blockmask) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) ((iocb->ki_pos + count) & mp->m_blockmask)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) unaligned_io = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) * We can't properly handle unaligned direct I/O to reflink
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) * files yet, as we can't unshare a partial block.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) if (xfs_is_cow_inode(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) trace_xfs_reflink_bounce_dio_write(ip, iocb->ki_pos, count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) return -ENOTBLK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) iolock = XFS_IOLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) iolock = XFS_IOLOCK_SHARED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) if (iocb->ki_flags & IOCB_NOWAIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552) /* unaligned dio always waits, bail */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) if (unaligned_io)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) if (!xfs_ilock_nowait(ip, iolock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558) xfs_ilock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561) ret = xfs_file_aio_write_checks(iocb, from, &iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) count = iov_iter_count(from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) * If we are doing unaligned IO, we can't allow any other overlapping IO
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) * in-flight at the same time or we risk data corruption. Wait for all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) * other IO to drain before we submit. If the IO is aligned, demote the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) * iolock if we had to take the exclusive lock in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) * xfs_file_aio_write_checks() for other reasons.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) if (unaligned_io) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574) inode_dio_wait(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) } else if (iolock == XFS_IOLOCK_EXCL) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) xfs_ilock_demote(ip, XFS_IOLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) iolock = XFS_IOLOCK_SHARED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) trace_xfs_file_direct_write(ip, count, iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582) * If unaligned, this is the only IO in-flight. Wait on it before we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) * release the iolock to prevent subsequent overlapping IO.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) ret = iomap_dio_rw(iocb, from, &xfs_direct_write_iomap_ops,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) &xfs_dio_write_ops,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) is_sync_kiocb(iocb) || unaligned_io);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) * No fallback to buffered IO after short writes for XFS, direct I/O
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593) * will either complete fully or return an error.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) ASSERT(ret < 0 || ret == count);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) static noinline ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) xfs_file_dax_write(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) struct iov_iter *from)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604) struct inode *inode = iocb->ki_filp->f_mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) int iolock = XFS_IOLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) ssize_t ret, error = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) size_t count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609) loff_t pos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) if (iocb->ki_flags & IOCB_NOWAIT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) if (!xfs_ilock_nowait(ip, iolock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613) return -EAGAIN;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) xfs_ilock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) ret = xfs_file_aio_write_checks(iocb, from, &iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622) pos = iocb->ki_pos;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) count = iov_iter_count(from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) trace_xfs_file_dax_write(ip, count, pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) ret = dax_iomap_rw(iocb, from, &xfs_direct_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) if (ret > 0 && iocb->ki_pos > i_size_read(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) i_size_write(inode, iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) error = xfs_setfilesize(ip, pos, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) if (ret > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) /* Handle various SYNC-type writes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) ret = generic_write_sync(iocb, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) xfs_file_buffered_aio_write(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) struct iov_iter *from)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) struct file *file = iocb->ki_filp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651) struct address_space *mapping = file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) struct inode *inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) ssize_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) int enospc = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) int iolock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658) if (iocb->ki_flags & IOCB_NOWAIT)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661) write_retry:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) iolock = XFS_IOLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) xfs_ilock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) ret = xfs_file_aio_write_checks(iocb, from, &iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) /* We can write back this queue in page reclaim */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) current->backing_dev_info = inode_to_bdi(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672) trace_xfs_file_buffered_write(ip, iov_iter_count(from), iocb->ki_pos);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) ret = iomap_file_buffered_write(iocb, from,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) &xfs_buffered_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) if (likely(ret >= 0))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) iocb->ki_pos += ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) * If we hit a space limit, try to free up some lingering preallocated
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680) * space before returning an error. In the case of ENOSPC, first try to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) * write back all dirty inodes to free up some of the excess reserved
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) * metadata space. This reduces the chances that the eofblocks scan
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683) * waits on dirty mappings. Since xfs_flush_inodes() is serialized, this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) * also behaves as a filter to prevent too many eofblocks scans from
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) * running at the same time.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) if (ret == -EDQUOT && !enospc) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) enospc = xfs_inode_free_quota_eofblocks(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690) if (enospc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) goto write_retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) enospc = xfs_inode_free_quota_cowblocks(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) if (enospc)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) goto write_retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) iolock = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) } else if (ret == -ENOSPC && !enospc) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) struct xfs_eofblocks eofb = {0};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) enospc = 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700) xfs_flush_inodes(ip->i_mount);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703) eofb.eof_flags = XFS_EOF_FLAGS_SYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) xfs_icache_free_eofblocks(ip->i_mount, &eofb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) xfs_icache_free_cowblocks(ip->i_mount, &eofb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706) goto write_retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) current->backing_dev_info = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) if (iolock)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) if (ret > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) XFS_STATS_ADD(ip->i_mount, xs_write_bytes, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) /* Handle various SYNC-type writes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) ret = generic_write_sync(iocb, ret);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) STATIC ssize_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) xfs_file_write_iter(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) struct kiocb *iocb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) struct iov_iter *from)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) struct file *file = iocb->ki_filp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) struct address_space *mapping = file->f_mapping;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729) struct inode *inode = mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) ssize_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) size_t ocount = iov_iter_count(from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) XFS_STATS_INC(ip->i_mount, xs_write_calls);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736) if (ocount == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) if (XFS_FORCED_SHUTDOWN(ip->i_mount))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) if (IS_DAX(inode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743) return xfs_file_dax_write(iocb, from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) if (iocb->ki_flags & IOCB_DIRECT) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) * Allow a directio write to fall back to a buffered
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) * write *only* in the case that we're doing a reflink
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) * CoW. In all other directio scenarios we do not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750) * allow an operation to fall back to buffered mode.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) ret = xfs_file_dio_aio_write(iocb, from);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) if (ret != -ENOTBLK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) return xfs_file_buffered_aio_write(iocb, from);
^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) static void
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) xfs_wait_dax_page(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) struct inode *inode)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) schedule();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) xfs_break_dax_layouts(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) bool *retry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) struct page *page;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778) ASSERT(xfs_isilocked(XFS_I(inode), XFS_MMAPLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) page = dax_layout_busy_page(inode->i_mapping);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) if (!page)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784) *retry = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) return ___wait_var_event(&page->_refcount,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) atomic_read(&page->_refcount) == 1, TASK_INTERRUPTIBLE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) 0, 0, xfs_wait_dax_page(inode));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790) int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) xfs_break_layouts(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) uint *iolock,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) enum layout_break_reason reason)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796) bool retry;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) ASSERT(xfs_isilocked(XFS_I(inode), XFS_IOLOCK_SHARED|XFS_IOLOCK_EXCL));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) do {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) retry = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) switch (reason) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804) case BREAK_UNMAP:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) error = xfs_break_dax_layouts(inode, &retry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) if (error || retry)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) /* fall through */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) case BREAK_WRITE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) error = xfs_break_leased_layouts(inode, iolock, &retry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) WARN_ON_ONCE(1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816) } while (error == 0 && retry);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821) #define XFS_FALLOC_FL_SUPPORTED \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) (FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE | \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) FALLOC_FL_COLLAPSE_RANGE | FALLOC_FL_ZERO_RANGE | \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) FALLOC_FL_INSERT_RANGE | FALLOC_FL_UNSHARE_RANGE)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) STATIC long
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) xfs_file_fallocate(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) int mode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) loff_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) loff_t len)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) struct inode *inode = file_inode(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) long error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) enum xfs_prealloc_flags flags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837) uint iolock = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) loff_t new_size = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) bool do_file_insert = false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) if (!S_ISREG(inode->i_mode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) if (mode & ~XFS_FALLOC_FL_SUPPORTED)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) xfs_ilock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) error = xfs_break_layouts(inode, &iolock, BREAK_UNMAP);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) * Must wait for all AIO to complete before we continue as AIO can
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) * change the file size on completion without holding any locks we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) * currently hold. We must do this first because AIO can update both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) * the on disk and in memory inode sizes, and the operations that follow
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) * require the in-memory size to be fully up-to-date.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) inode_dio_wait(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861) * Now AIO and DIO has drained we flush and (if necessary) invalidate
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) * the cached range over the first operation we are about to run.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) * We care about zero and collapse here because they both run a hole
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865) * punch over the range first. Because that can zero data, and the range
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) * of invalidation for the shift operations is much larger, we still do
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) * the required flush for collapse in xfs_prepare_shift().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) * Insert has the same range requirements as collapse, and we extend the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870) * file first which can zero data. Hence insert has the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) * flush/invalidate requirements as collapse and so they are both
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) * handled at the right time by xfs_prepare_shift().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) if (mode & (FALLOC_FL_PUNCH_HOLE | FALLOC_FL_ZERO_RANGE |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875) FALLOC_FL_COLLAPSE_RANGE)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) error = xfs_flush_unmap_range(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 877) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 878) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 879) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 880)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 881) if (mode & FALLOC_FL_PUNCH_HOLE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 882) error = xfs_free_file_space(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 883) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 884) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 885) } else if (mode & FALLOC_FL_COLLAPSE_RANGE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 886) if (!xfs_is_falloc_aligned(ip, offset, len)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 887) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 888) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 889) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 890)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 891) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 892) * There is no need to overlap collapse range with EOF,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 893) * in which case it is effectively a truncate operation
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 894) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 895) if (offset + len >= i_size_read(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 896) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 897) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 898) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 899)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 900) new_size = i_size_read(inode) - len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 901)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 902) error = xfs_collapse_file_space(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 903) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 904) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 905) } else if (mode & FALLOC_FL_INSERT_RANGE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 906) loff_t isize = i_size_read(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 907)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 908) if (!xfs_is_falloc_aligned(ip, offset, len)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 909) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 910) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 911) }
^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) * New inode size must not exceed ->s_maxbytes, accounting for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 915) * possible signed overflow.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 916) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 917) if (inode->i_sb->s_maxbytes - isize < len) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 918) error = -EFBIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 919) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 920) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 921) new_size = isize + len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 922)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 923) /* Offset should be less than i_size */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 924) if (offset >= isize) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 925) error = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 926) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 927) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 928) do_file_insert = true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 929) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 930) flags |= XFS_PREALLOC_SET;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 931)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 932) if (!(mode & FALLOC_FL_KEEP_SIZE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 933) offset + len > i_size_read(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 934) new_size = offset + len;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 935) error = inode_newsize_ok(inode, new_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 936) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 937) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 938) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 939)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 940) if (mode & FALLOC_FL_ZERO_RANGE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 941) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 942) * Punch a hole and prealloc the range. We use a hole
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 943) * punch rather than unwritten extent conversion for two
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 944) * reasons:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 945) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 946) * 1.) Hole punch handles partial block zeroing for us.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 947) * 2.) If prealloc returns ENOSPC, the file range is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 948) * still zero-valued by virtue of the hole punch.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 949) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 950) unsigned int blksize = i_blocksize(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 951)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 952) trace_xfs_zero_file_space(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 953)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 954) error = xfs_free_file_space(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 955) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 956) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 957)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 958) len = round_up(offset + len, blksize) -
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 959) round_down(offset, blksize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 960) offset = round_down(offset, blksize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 961) } else if (mode & FALLOC_FL_UNSHARE_RANGE) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 962) error = xfs_reflink_unshare(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 963) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 964) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 965) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 966) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 967) * If always_cow mode we can't use preallocations and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 968) * thus should not create them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 969) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 970) if (xfs_is_always_cow_inode(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 971) error = -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 972) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 973) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 974) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 975)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 976) if (!xfs_is_always_cow_inode(ip)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 977) error = xfs_alloc_file_space(ip, offset, len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 978) XFS_BMAPI_PREALLOC);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 979) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 980) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 981) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 982) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 983)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 984) if (file->f_flags & O_DSYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 985) flags |= XFS_PREALLOC_SYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 986)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 987) error = xfs_update_prealloc_flags(ip, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 988) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 989) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 990)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 991) /* Change file size if needed */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 992) if (new_size) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 993) struct iattr iattr;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 994)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 995) iattr.ia_valid = ATTR_SIZE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 996) iattr.ia_size = new_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 997) error = xfs_vn_setattr_size(file_dentry(file), &iattr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 998) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 999) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1000) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1001)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1002) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1003) * Perform hole insertion now that the file size has been
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1004) * updated so that if we crash during the operation we don't
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1005) * leave shifted extents past EOF and hence losing access to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1006) * the data that is contained within them.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1007) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1008) if (do_file_insert)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1009) error = xfs_insert_file_space(ip, offset, len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1010)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1011) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1012) xfs_iunlock(ip, iolock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1013) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1014) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1015)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1016) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1017) xfs_file_fadvise(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1018) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1019) loff_t start,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1020) loff_t end,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1021) int advice)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1022) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1023) struct xfs_inode *ip = XFS_I(file_inode(file));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1024) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1025) int lockflags = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1026)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1027) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1028) * Operations creating pages in page cache need protection from hole
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1029) * punching and similar ops
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1030) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1031) if (advice == POSIX_FADV_WILLNEED) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1032) lockflags = XFS_IOLOCK_SHARED;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1033) xfs_ilock(ip, lockflags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1034) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1035) ret = generic_fadvise(file, start, end, advice);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1036) if (lockflags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1037) xfs_iunlock(ip, lockflags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1038) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1039) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1040)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1041) /* Does this file, inode, or mount want synchronous writes? */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1042) static inline bool xfs_file_sync_writes(struct file *filp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1043) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1044) struct xfs_inode *ip = XFS_I(file_inode(filp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1045)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1046) if (ip->i_mount->m_flags & XFS_MOUNT_WSYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1047) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1048) if (filp->f_flags & (__O_SYNC | O_DSYNC))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1049) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1050) if (IS_SYNC(file_inode(filp)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1051) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1052)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1053) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1054) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1055)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1056) STATIC loff_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1057) xfs_file_remap_range(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1058) struct file *file_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1059) loff_t pos_in,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1060) struct file *file_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1061) loff_t pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1062) loff_t len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1063) unsigned int remap_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1064) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1065) struct inode *inode_in = file_inode(file_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1066) struct xfs_inode *src = XFS_I(inode_in);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1067) struct inode *inode_out = file_inode(file_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1068) struct xfs_inode *dest = XFS_I(inode_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1069) struct xfs_mount *mp = src->i_mount;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1070) loff_t remapped = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1071) xfs_extlen_t cowextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1072) int ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1073)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1074) if (remap_flags & ~(REMAP_FILE_DEDUP | REMAP_FILE_ADVISORY))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1075) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1076)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1077) if (!xfs_sb_version_hasreflink(&mp->m_sb))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1078) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1079)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1080) if (XFS_FORCED_SHUTDOWN(mp))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1081) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1082)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1083) /* Prepare and then clone file data. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1084) ret = xfs_reflink_remap_prep(file_in, pos_in, file_out, pos_out,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1085) &len, remap_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1086) if (ret || len == 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1087) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1088)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1089) trace_xfs_reflink_remap_range(src, pos_in, len, dest, pos_out);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1090)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1091) ret = xfs_reflink_remap_blocks(src, pos_in, dest, pos_out, len,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1092) &remapped);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1093) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1094) goto out_unlock;
^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) * Carry the cowextsize hint from src to dest if we're sharing the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1098) * entire source file to the entire destination file, the source file
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1099) * has a cowextsize hint, and the destination file does not.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1100) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1101) cowextsize = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1102) if (pos_in == 0 && len == i_size_read(inode_in) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1103) (src->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1104) pos_out == 0 && len >= i_size_read(inode_out) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1105) !(dest->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1106) cowextsize = src->i_d.di_cowextsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1107)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1108) ret = xfs_reflink_update_dest(dest, pos_out + len, cowextsize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1109) remap_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1110) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1111) goto out_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1112)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1113) if (xfs_file_sync_writes(file_in) || xfs_file_sync_writes(file_out))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1114) xfs_log_force_inode(dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1115) out_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1116) xfs_iunlock2_io_mmap(src, dest);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1117) if (ret)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1118) trace_xfs_reflink_remap_range_error(dest, ret, _RET_IP_);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1119) return remapped > 0 ? remapped : ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1120) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1121)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1122) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1123) xfs_file_open(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1124) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1125) struct file *file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1126) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1127) if (!(file->f_flags & O_LARGEFILE) && i_size_read(inode) > MAX_NON_LFS)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1128) return -EFBIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1129) if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1130) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1131) file->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1132) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1133) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1134)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1135) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1136) xfs_dir_open(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1137) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1138) struct file *file)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1139) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1140) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1141) int mode;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1142) int error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1143)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1144) error = xfs_file_open(inode, file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1145) if (error)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1146) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1147)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1148) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1149) * If there are any blocks, read-ahead block 0 as we're almost
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1150) * certain to have the next operation be a read there.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1151) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1152) mode = xfs_ilock_data_map_shared(ip);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1153) if (ip->i_df.if_nextents > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1154) error = xfs_dir3_data_readahead(ip, 0, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1155) xfs_iunlock(ip, mode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1156) return error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1157) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1158)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1159) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1160) xfs_file_release(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1161) struct inode *inode,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1162) struct file *filp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1163) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1164) return xfs_release(XFS_I(inode));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1165) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1166)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1167) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1168) xfs_file_readdir(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1169) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1170) struct dir_context *ctx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1171) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1172) struct inode *inode = file_inode(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1173) xfs_inode_t *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1174) size_t bufsize;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1175)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1176) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1177) * The Linux API doesn't pass down the total size of the buffer
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1178) * we read into down to the filesystem. With the filldir concept
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1179) * it's not needed for correct information, but the XFS dir2 leaf
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1180) * code wants an estimate of the buffer size to calculate it's
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1181) * readahead window and size the buffers used for mapping to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1182) * physical blocks.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1183) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1184) * Try to give it an estimate that's good enough, maybe at some
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1185) * point we can change the ->readdir prototype to include the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1186) * buffer size. For now we use the current glibc buffer size.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1187) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1188) bufsize = (size_t)min_t(loff_t, XFS_READDIR_BUFSIZE, ip->i_d.di_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1189)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1190) return xfs_readdir(NULL, ip, ctx, bufsize);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1191) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1192)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1193) STATIC loff_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1194) xfs_file_llseek(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1195) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1196) loff_t offset,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1197) int whence)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1198) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1199) struct inode *inode = file->f_mapping->host;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1200)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1201) if (XFS_FORCED_SHUTDOWN(XFS_I(inode)->i_mount))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1202) return -EIO;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1203)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1204) switch (whence) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1205) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1206) return generic_file_llseek(file, offset, whence);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1207) case SEEK_HOLE:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1208) offset = iomap_seek_hole(inode, offset, &xfs_seek_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1209) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1210) case SEEK_DATA:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1211) offset = iomap_seek_data(inode, offset, &xfs_seek_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1212) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1213) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1214)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1215) if (offset < 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1216) return offset;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1217) return vfs_setpos(file, offset, inode->i_sb->s_maxbytes);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1218) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1220) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1221) * Locking for serialisation of IO during page faults. This results in a lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1222) * ordering of:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1223) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1224) * mmap_lock (MM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1225) * sb_start_pagefault(vfs, freeze)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1226) * i_mmaplock (XFS - truncate serialisation)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1227) * page_lock (MM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1228) * i_lock (XFS - extent map serialisation)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1229) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1230) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1231) __xfs_filemap_fault(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1232) struct vm_fault *vmf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1233) enum page_entry_size pe_size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1234) bool write_fault)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1235) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1236) struct inode *inode = file_inode(vmf->vma->vm_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1237) struct xfs_inode *ip = XFS_I(inode);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1238) vm_fault_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1239)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1240) trace_xfs_filemap_fault(ip, pe_size, write_fault);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1242) if (write_fault) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1243) sb_start_pagefault(inode->i_sb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1244) file_update_time(vmf->vma->vm_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1245) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1246)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1247) xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1248) if (IS_DAX(inode)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1249) pfn_t pfn;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1250)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1251) ret = dax_iomap_fault(vmf, pe_size, &pfn, NULL,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1252) (write_fault && !vmf->cow_page) ?
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1253) &xfs_direct_write_iomap_ops :
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1254) &xfs_read_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1255) if (ret & VM_FAULT_NEEDDSYNC)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1256) ret = dax_finish_sync_fault(vmf, pe_size, pfn);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1257) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1258) if (write_fault)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1259) ret = iomap_page_mkwrite(vmf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1260) &xfs_buffered_write_iomap_ops);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1261) else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1262) ret = filemap_fault(vmf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1263) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1264) xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1265)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1266) if (write_fault)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1267) sb_end_pagefault(inode->i_sb);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1268) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1269) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1270)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1271) static inline bool
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1272) xfs_is_write_fault(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1273) struct vm_fault *vmf)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1274) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1275) return (vmf->flags & FAULT_FLAG_WRITE) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1276) (vmf->vma->vm_flags & VM_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1277) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1278)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1279) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1280) xfs_filemap_fault(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1281) struct vm_fault *vmf)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1282) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1283) /* DAX can shortcut the normal fault path on write faults! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1284) return __xfs_filemap_fault(vmf, PE_SIZE_PTE,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1285) IS_DAX(file_inode(vmf->vma->vm_file)) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1286) xfs_is_write_fault(vmf));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1287) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1288)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1289) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1290) xfs_filemap_huge_fault(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1291) struct vm_fault *vmf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1292) enum page_entry_size pe_size)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1293) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1294) if (!IS_DAX(file_inode(vmf->vma->vm_file)))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1295) return VM_FAULT_FALLBACK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1296)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1297) /* DAX can shortcut the normal fault path on write faults! */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1298) return __xfs_filemap_fault(vmf, pe_size,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1299) xfs_is_write_fault(vmf));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1300) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1301)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1302) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1303) xfs_filemap_page_mkwrite(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1304) struct vm_fault *vmf)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1305) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1306) return __xfs_filemap_fault(vmf, PE_SIZE_PTE, true);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1307) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1308)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1309) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1310) * pfn_mkwrite was originally intended to ensure we capture time stamp updates
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1311) * on write faults. In reality, it needs to serialise against truncate and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1312) * prepare memory for writing so handle is as standard write fault.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1313) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1314) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1315) xfs_filemap_pfn_mkwrite(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1316) struct vm_fault *vmf)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1317) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1318)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1319) return __xfs_filemap_fault(vmf, PE_SIZE_PTE, true);
^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) static vm_fault_t
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1323) xfs_filemap_map_pages(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1324) struct vm_fault *vmf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1325) pgoff_t start_pgoff,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1326) pgoff_t end_pgoff)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1327) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1328) struct inode *inode = file_inode(vmf->vma->vm_file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1329) vm_fault_t ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1331) xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1332) ret = filemap_map_pages(vmf, start_pgoff, end_pgoff);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1333) xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1334) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1335) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1336)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1337) static const struct vm_operations_struct xfs_file_vm_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1338) .fault = xfs_filemap_fault,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1339) .huge_fault = xfs_filemap_huge_fault,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1340) .map_pages = xfs_filemap_map_pages,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1341) .page_mkwrite = xfs_filemap_page_mkwrite,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1342) .pfn_mkwrite = xfs_filemap_pfn_mkwrite,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1343) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1344)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1345) STATIC int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1346) xfs_file_mmap(
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1347) struct file *file,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1348) struct vm_area_struct *vma)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1349) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1350) struct inode *inode = file_inode(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1351) struct xfs_buftarg *target = xfs_inode_buftarg(XFS_I(inode));
^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) * We don't support synchronous mappings for non-DAX files and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1355) * for DAX files if underneath dax_device is not synchronous.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1356) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1357) if (!daxdev_mapping_supported(vma, target->bt_daxdev))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1358) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1359)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1360) file_accessed(file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1361) vma->vm_ops = &xfs_file_vm_ops;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1362) if (IS_DAX(inode))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1363) vma->vm_flags |= VM_HUGEPAGE;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1364) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1365) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1366)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1367) const struct file_operations xfs_file_operations = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1368) .llseek = xfs_file_llseek,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1369) .read_iter = xfs_file_read_iter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1370) .write_iter = xfs_file_write_iter,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1371) .splice_read = generic_file_splice_read,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1372) .splice_write = iter_file_splice_write,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1373) .iopoll = iomap_dio_iopoll,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1374) .unlocked_ioctl = xfs_file_ioctl,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1375) #ifdef CONFIG_COMPAT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1376) .compat_ioctl = xfs_file_compat_ioctl,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1377) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1378) .mmap = xfs_file_mmap,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1379) .mmap_supported_flags = MAP_SYNC,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1380) .open = xfs_file_open,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1381) .release = xfs_file_release,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1382) .fsync = xfs_file_fsync,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1383) .get_unmapped_area = thp_get_unmapped_area,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1384) .fallocate = xfs_file_fallocate,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1385) .fadvise = xfs_file_fadvise,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1386) .remap_file_range = xfs_file_remap_range,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1387) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1388)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1389) const struct file_operations xfs_dir_file_operations = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1390) .open = xfs_dir_open,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1391) .read = generic_read_dir,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1392) .iterate_shared = xfs_file_readdir,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1393) .llseek = generic_file_llseek,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1394) .unlocked_ioctl = xfs_file_ioctl,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1395) #ifdef CONFIG_COMPAT
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1396) .compat_ioctl = xfs_file_compat_ioctl,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1397) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1398) .fsync = xfs_dir_fsync,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1399) };