ssize_t /* bytes read, or (-) error */ xfs_read( bhv_desc_t *bdp, struct kiocb *iocb, const struct iovec *iovp, unsigned int segs, loff_t *offset, int ioflags, cred_t *credp) { struct file *file = iocb->ki_filp; struct inode *inode = file->f_mapping->host; size_t size = 0; ssize_t ret; xfs_fsize_t n; xfs_inode_t *ip; xfs_mount_t *mp; bhv_vnode_t *vp; unsigned long seg; ip = XFS_BHVTOI(bdp); vp = BHV_TO_VNODE(bdp); mp = ip->i_mount; XFS_STATS_INC(xs_read_calls); /* START copy & waste from filemap.c */ for (seg = 0; seg < segs; seg++) { const struct iovec *iv = &iovp[seg]; /* * If any segment has a negative length, or the cumulative * length ever wraps negative then return -EINVAL. */ size += iv->iov_len; if (unlikely((ssize_t)(size|iv->iov_len) < 0)) return XFS_ERROR(-EINVAL); } /* END copy & waste from filemap.c */ if (unlikely(ioflags & IO_ISDIRECT)) { xfs_buftarg_t *target = (ip->i_d.di_flags & XFS_DIFLAG_REALTIME) ? mp->m_rtdev_targp : mp->m_ddev_targp; if ((*offset & target->bt_smask) || (size & target->bt_smask)) { if (*offset == ip->i_d.di_size) { return (0); } return -XFS_ERROR(EINVAL); } } n = XFS_MAXIOFFSET(mp) - *offset; if ((n <= 0) || (size == 0)) return 0; if (n < size) size = n; if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; if (unlikely(ioflags & IO_ISDIRECT)) mutex_lock(&inode->i_mutex); xfs_ilock(ip, XFS_IOLOCK_SHARED); if (DM_EVENT_ENABLED(vp->v_vfsp, ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) { bhv_vrwlock_t locktype = VRWLOCK_READ; int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags); ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, BHV_TO_VNODE(bdp), *offset, size, dmflags, &locktype); if (ret) { xfs_iunlock(ip, XFS_IOLOCK_SHARED); if (unlikely(ioflags & IO_ISDIRECT)) mutex_unlock(&inode->i_mutex); return ret; } } if (unlikely(ioflags & IO_ISDIRECT)) { if (VN_CACHED(vp)) bhv_vop_flushinval_pages(vp, ctooff(offtoct(*offset)), -1, FI_REMAPF_LOCKED); mutex_unlock(&inode->i_mutex); } xfs_rw_enter_trace(XFS_READ_ENTER, &ip->i_iocore, (void *)iovp, segs, *offset, ioflags); iocb->ki_pos = *offset; ret = generic_file_aio_read(iocb, iovp, segs, *offset); if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO)) ret = wait_on_sync_kiocb(iocb); if (ret > 0) XFS_STATS_ADD(xs_read_bytes, ret); xfs_iunlock(ip, XFS_IOLOCK_SHARED); return ret; }
ssize_t /* bytes written, or (-) error */ xfs_write( bhv_desc_t *bdp, struct kiocb *iocb, const struct iovec *iovp, unsigned int nsegs, loff_t *offset, int ioflags, cred_t *credp) { struct file *file = iocb->ki_filp; struct address_space *mapping = file->f_mapping; struct inode *inode = mapping->host; unsigned long segs = nsegs; xfs_inode_t *xip; xfs_mount_t *mp; ssize_t ret = 0, error = 0; xfs_fsize_t isize, new_size; xfs_iocore_t *io; bhv_vnode_t *vp; unsigned long seg; int iolock; int eventsent = 0; bhv_vrwlock_t locktype; size_t ocount = 0, count; loff_t pos; int need_i_mutex = 1, need_flush = 0; XFS_STATS_INC(xs_write_calls); vp = BHV_TO_VNODE(bdp); xip = XFS_BHVTOI(bdp); for (seg = 0; seg < segs; seg++) { const struct iovec *iv = &iovp[seg]; /* * If any segment has a negative length, or the cumulative * length ever wraps negative then return -EINVAL. */ ocount += iv->iov_len; if (unlikely((ssize_t)(ocount|iv->iov_len) < 0)) return -EINVAL; if (access_ok(VERIFY_READ, iv->iov_base, iv->iov_len)) continue; if (seg == 0) return -EFAULT; segs = seg; ocount -= iv->iov_len; /* This segment is no good */ break; } count = ocount; pos = *offset; if (count == 0) return 0; io = &xip->i_iocore; mp = io->io_mount; vfs_wait_for_freeze(vp->v_vfsp, SB_FREEZE_WRITE); if (XFS_FORCED_SHUTDOWN(mp)) return -EIO; if (ioflags & IO_ISDIRECT) { xfs_buftarg_t *target = (xip->i_d.di_flags & XFS_DIFLAG_REALTIME) ? mp->m_rtdev_targp : mp->m_ddev_targp; if ((pos & target->bt_smask) || (count & target->bt_smask)) return XFS_ERROR(-EINVAL); if (!VN_CACHED(vp) && pos < i_size_read(inode)) need_i_mutex = 0; if (VN_CACHED(vp)) need_flush = 1; } relock: if (need_i_mutex) { iolock = XFS_IOLOCK_EXCL; locktype = VRWLOCK_WRITE; mutex_lock(&inode->i_mutex); } else { iolock = XFS_IOLOCK_SHARED; locktype = VRWLOCK_WRITE_DIRECT; } xfs_ilock(xip, XFS_ILOCK_EXCL|iolock); isize = i_size_read(inode); if (file->f_flags & O_APPEND) *offset = isize; start: error = -generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode)); if (error) { xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock); goto out_unlock_mutex; } new_size = pos + count; if (new_size > isize) io->io_new_size = new_size; if ((DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS) && !eventsent)) { loff_t savedsize = pos; int dmflags = FILP_DELAY_FLAG(file); if (need_i_mutex) dmflags |= DM_FLAGS_IMUX; xfs_iunlock(xip, XFS_ILOCK_EXCL); error = XFS_SEND_DATA(xip->i_mount, DM_EVENT_WRITE, vp, pos, count, dmflags, &locktype); if (error) { xfs_iunlock(xip, iolock); goto out_unlock_mutex; } xfs_ilock(xip, XFS_ILOCK_EXCL); eventsent = 1; /* * The iolock was dropped and reacquired in XFS_SEND_DATA * so we have to recheck the size when appending. * We will only "goto start;" once, since having sent the * event prevents another call to XFS_SEND_DATA, which is * what allows the size to change in the first place. */ if ((file->f_flags & O_APPEND) && savedsize != isize) { pos = isize = xip->i_d.di_size; goto start; } } if (likely(!(ioflags & IO_INVIS))) { file_update_time(file); xfs_ichgtime_fast(xip, inode, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG); } /* * If the offset is beyond the size of the file, we have a couple * of things to do. First, if there is already space allocated * we need to either create holes or zero the disk or ... * * If there is a page where the previous size lands, we need * to zero it out up to the new size. */ if (pos > isize) { error = xfs_zero_eof(BHV_TO_VNODE(bdp), io, pos, isize); if (error) { xfs_iunlock(xip, XFS_ILOCK_EXCL|iolock); goto out_unlock_mutex; } } xfs_iunlock(xip, XFS_ILOCK_EXCL); /* * If we're writing the file then make sure to clear the * setuid and setgid bits if the process is not being run * by root. This keeps people from modifying setuid and * setgid binaries. */ if (((xip->i_d.di_mode & S_ISUID) || ((xip->i_d.di_mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP))) && !capable(CAP_FSETID)) { error = xfs_write_clear_setuid(xip); if (likely(!error)) error = -remove_suid(file->f_path.dentry); if (unlikely(error)) { xfs_iunlock(xip, iolock); goto out_unlock_mutex; } } retry: /* We can write back this queue in page reclaim */ current->backing_dev_info = mapping->backing_dev_info; if ((ioflags & IO_ISDIRECT)) { if (need_flush) { xfs_inval_cached_trace(io, pos, -1, ctooff(offtoct(pos)), -1); bhv_vop_flushinval_pages(vp, ctooff(offtoct(pos)), -1, FI_REMAPF_LOCKED); } if (need_i_mutex) { /* demote the lock now the cached pages are gone */ XFS_ILOCK_DEMOTE(mp, io, XFS_IOLOCK_EXCL); mutex_unlock(&inode->i_mutex); iolock = XFS_IOLOCK_SHARED; locktype = VRWLOCK_WRITE_DIRECT; need_i_mutex = 0; } xfs_rw_enter_trace(XFS_DIOWR_ENTER, io, (void *)iovp, segs, *offset, ioflags); ret = generic_file_direct_write(iocb, iovp, &segs, pos, offset, count, ocount); /* * direct-io write to a hole: fall through to buffered I/O * for completing the rest of the request. */ if (ret >= 0 && ret != count) { XFS_STATS_ADD(xs_write_bytes, ret); pos += ret; count -= ret; need_i_mutex = 1; ioflags &= ~IO_ISDIRECT; xfs_iunlock(xip, iolock); goto relock; } } else { xfs_rw_enter_trace(XFS_WRITE_ENTER, io, (void *)iovp, segs, *offset, ioflags); ret = generic_file_buffered_write(iocb, iovp, segs, pos, offset, count, ret); } current->backing_dev_info = NULL; if (ret == -EIOCBQUEUED && !(ioflags & IO_ISAIO)) ret = wait_on_sync_kiocb(iocb); if ((ret == -ENOSPC) && DM_EVENT_ENABLED(vp->v_vfsp, xip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) { xfs_rwunlock(bdp, locktype); if (need_i_mutex) mutex_unlock(&inode->i_mutex); error = XFS_SEND_NAMESP(xip->i_mount, DM_EVENT_NOSPACE, vp, DM_RIGHT_NULL, vp, DM_RIGHT_NULL, NULL, NULL, 0, 0, 0); /* Delay flag intentionally unused */ if (error) goto out_nounlocks; if (need_i_mutex) mutex_lock(&inode->i_mutex); xfs_rwlock(bdp, locktype); pos = xip->i_d.di_size; ret = 0; goto retry; } isize = i_size_read(inode); if (unlikely(ret < 0 && ret != -EFAULT && *offset > isize)) *offset = isize; if (*offset > xip->i_d.di_size) { xfs_ilock(xip, XFS_ILOCK_EXCL); if (*offset > xip->i_d.di_size) { xip->i_d.di_size = *offset; i_size_write(inode, *offset); xip->i_update_core = 1; xip->i_update_size = 1; } xfs_iunlock(xip, XFS_ILOCK_EXCL); } error = -ret; if (ret <= 0) goto out_unlock_internal; XFS_STATS_ADD(xs_write_bytes, ret); /* Handle various SYNC-type writes */ if ((file->f_flags & O_SYNC) || IS_SYNC(inode)) { error = xfs_write_sync_logforce(mp, xip); if (error) goto out_unlock_internal; xfs_rwunlock(bdp, locktype); if (need_i_mutex) mutex_unlock(&inode->i_mutex); error = sync_page_range(inode, mapping, pos, ret); if (!error) error = ret; return error; } out_unlock_internal: xfs_rwunlock(bdp, locktype); out_unlock_mutex: if (need_i_mutex) mutex_unlock(&inode->i_mutex); out_nounlocks: return -error; }
int xfs_swap_extents( xfs_inode_t *ip, xfs_inode_t *tip, xfs_swapext_t *sxp) { xfs_mount_t *mp; xfs_inode_t *ips[2]; xfs_trans_t *tp; xfs_bstat_t *sbp = &sxp->sx_stat; bhv_vnode_t *vp, *tvp; xfs_ifork_t *tempifp, *ifp, *tifp; int ilf_fields, tilf_fields; static uint lock_flags = XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL; int error = 0; int aforkblks = 0; int taforkblks = 0; __uint64_t tmp; char locked = 0; mp = ip->i_mount; tempifp = kmem_alloc(sizeof(xfs_ifork_t), KM_MAYFAIL); if (!tempifp) { error = XFS_ERROR(ENOMEM); goto error0; } sbp = &sxp->sx_stat; vp = XFS_ITOV(ip); tvp = XFS_ITOV(tip); /* Lock in i_ino order */ if (ip->i_ino < tip->i_ino) { ips[0] = ip; ips[1] = tip; } else { ips[0] = tip; ips[1] = ip; } xfs_lock_inodes(ips, 2, 0, lock_flags); locked = 1; /* Check permissions */ error = xfs_iaccess(ip, S_IWUSR, NULL); if (error) goto error0; error = xfs_iaccess(tip, S_IWUSR, NULL); if (error) goto error0; /* Verify that both files have the same format */ if ((ip->i_d.di_mode & S_IFMT) != (tip->i_d.di_mode & S_IFMT)) { error = XFS_ERROR(EINVAL); goto error0; } /* Verify both files are either real-time or non-realtime */ if ((ip->i_d.di_flags & XFS_DIFLAG_REALTIME) != (tip->i_d.di_flags & XFS_DIFLAG_REALTIME)) { error = XFS_ERROR(EINVAL); goto error0; } /* Should never get a local format */ if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL || tip->i_d.di_format == XFS_DINODE_FMT_LOCAL) { error = XFS_ERROR(EINVAL); goto error0; } if (VN_CACHED(tvp) != 0) { xfs_inval_cached_trace(&tip->i_iocore, 0, -1, 0, -1); bhv_vop_flushinval_pages(tvp, 0, -1, FI_REMAPF_LOCKED); } /* Verify O_DIRECT for ftmp */ if (VN_CACHED(tvp) != 0) { error = XFS_ERROR(EINVAL); goto error0; } /* Verify all data are being swapped */ if (sxp->sx_offset != 0 || sxp->sx_length != ip->i_d.di_size || sxp->sx_length != tip->i_d.di_size) { error = XFS_ERROR(EFAULT); goto error0; } /* * If the target has extended attributes, the tmp file * must also in order to ensure the correct data fork * format. */ if ( XFS_IFORK_Q(ip) != XFS_IFORK_Q(tip) ) { error = XFS_ERROR(EINVAL); goto error0; } /* * Compare the current change & modify times with that * passed in. If they differ, we abort this swap. * This is the mechanism used to ensure the calling * process that the file was not changed out from * under it. */ if ((sbp->bs_ctime.tv_sec != ip->i_d.di_ctime.t_sec) || (sbp->bs_ctime.tv_nsec != ip->i_d.di_ctime.t_nsec) || (sbp->bs_mtime.tv_sec != ip->i_d.di_mtime.t_sec) || (sbp->bs_mtime.tv_nsec != ip->i_d.di_mtime.t_nsec)) { error = XFS_ERROR(EBUSY); goto error0; } /* We need to fail if the file is memory mapped. Once we have tossed * all existing pages, the page fault will have no option * but to go to the filesystem for pages. By making the page fault call * vop_read (or write in the case of autogrow) they block on the iolock * until we have switched the extents. */ if (VN_MAPPED(vp)) { error = XFS_ERROR(EBUSY); goto error0; } xfs_iunlock(ip, XFS_ILOCK_EXCL); xfs_iunlock(tip, XFS_ILOCK_EXCL); /* * There is a race condition here since we gave up the * ilock. However, the data fork will not change since * we have the iolock (locked for truncation too) so we * are safe. We don't really care if non-io related * fields change. */ bhv_vop_toss_pages(vp, 0, -1, FI_REMAPF); tp = xfs_trans_alloc(mp, XFS_TRANS_SWAPEXT); if ((error = xfs_trans_reserve(tp, 0, XFS_ICHANGE_LOG_RES(mp), 0, 0, 0))) { xfs_iunlock(ip, XFS_IOLOCK_EXCL); xfs_iunlock(tip, XFS_IOLOCK_EXCL); xfs_trans_cancel(tp, 0); locked = 0; goto error0; } xfs_lock_inodes(ips, 2, 0, XFS_ILOCK_EXCL); /* * Count the number of extended attribute blocks */ if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) && (ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) { error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &aforkblks); if (error) { xfs_trans_cancel(tp, 0); goto error0; } } if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) && (tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) { error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK, &taforkblks); if (error) { xfs_trans_cancel(tp, 0); goto error0; } } /* * Swap the data forks of the inodes */ ifp = &ip->i_df; tifp = &tip->i_df; *tempifp = *ifp; /* struct copy */ *ifp = *tifp; /* struct copy */ *tifp = *tempifp; /* struct copy */ /* * Fix the on-disk inode values */ tmp = (__uint64_t)ip->i_d.di_nblocks; ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks; tip->i_d.di_nblocks = tmp + taforkblks - aforkblks; tmp = (__uint64_t) ip->i_d.di_nextents; ip->i_d.di_nextents = tip->i_d.di_nextents; tip->i_d.di_nextents = tmp; tmp = (__uint64_t) ip->i_d.di_format; ip->i_d.di_format = tip->i_d.di_format; tip->i_d.di_format = tmp; ilf_fields = XFS_ILOG_CORE; switch(ip->i_d.di_format) { case XFS_DINODE_FMT_EXTENTS: /* If the extents fit in the inode, fix the * pointer. Otherwise it's already NULL or * pointing to the extent. */ if (ip->i_d.di_nextents <= XFS_INLINE_EXTS) { ifp->if_u1.if_extents = ifp->if_u2.if_inline_ext; } ilf_fields |= XFS_ILOG_DEXT; break; case XFS_DINODE_FMT_BTREE: ilf_fields |= XFS_ILOG_DBROOT; break; } tilf_fields = XFS_ILOG_CORE; switch(tip->i_d.di_format) { case XFS_DINODE_FMT_EXTENTS: /* If the extents fit in the inode, fix the * pointer. Otherwise it's already NULL or * pointing to the extent. */ if (tip->i_d.di_nextents <= XFS_INLINE_EXTS) { tifp->if_u1.if_extents = tifp->if_u2.if_inline_ext; } tilf_fields |= XFS_ILOG_DEXT; break; case XFS_DINODE_FMT_BTREE: tilf_fields |= XFS_ILOG_DBROOT; break; } /* * Increment vnode ref counts since xfs_trans_commit & * xfs_trans_cancel will both unlock the inodes and * decrement the associated ref counts. */ VN_HOLD(vp); VN_HOLD(tvp); xfs_trans_ijoin(tp, ip, lock_flags); xfs_trans_ijoin(tp, tip, lock_flags); xfs_trans_log_inode(tp, ip, ilf_fields); xfs_trans_log_inode(tp, tip, tilf_fields); /* * If this is a synchronous mount, make sure that the * transaction goes to disk before returning to the user. */ if (mp->m_flags & XFS_MOUNT_WSYNC) { xfs_trans_set_sync(tp); } error = xfs_trans_commit(tp, XFS_TRANS_SWAPEXT, NULL); locked = 0; error0: if (locked) { xfs_iunlock(ip, lock_flags); xfs_iunlock(tip, lock_flags); } if (tempifp != NULL) kmem_free(tempifp, sizeof(xfs_ifork_t)); return error; }