+/*
+ * Will look for holes and unwritten extents in the range starting at
+ * pos for count bytes (inclusive).
+ */
+static int ocfs2_check_range_for_holes(struct inode *inode, loff_t pos,
+ size_t count)
+{
+ int ret = 0;
+ unsigned int extent_flags;
+ u32 cpos, clusters, extent_len, phys_cpos;
+ struct super_block *sb = inode->i_sb;
+
+ cpos = pos >> OCFS2_SB(sb)->s_clustersize_bits;
+ clusters = ocfs2_clusters_for_bytes(sb, pos + count) - cpos;
+
+ while (clusters) {
+ ret = ocfs2_get_clusters(inode, cpos, &phys_cpos, &extent_len,
+ &extent_flags);
+ if (ret < 0) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ if (phys_cpos == 0 || (extent_flags & OCFS2_EXT_UNWRITTEN)) {
+ ret = 1;
+ break;
+ }
+
+ if (extent_len > clusters)
+ extent_len = clusters;
+
+ clusters -= extent_len;
+ cpos += extent_len;
+ }
+out:
+ return ret;
+}
+
+static int ocfs2_write_remove_suid(struct inode *inode)
+{
+ int ret;
+ struct buffer_head *bh = NULL;
+
+ ret = ocfs2_read_inode_block(inode, &bh);
+ if (ret < 0) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = __ocfs2_write_remove_suid(inode, bh);
+out:
+ brelse(bh);
+ return ret;
+}
+
+/*
+ * Allocate enough extents to cover the region starting at byte offset
+ * start for len bytes. Existing extents are skipped, any extents
+ * added are marked as "unwritten".
+ */
+static int ocfs2_allocate_unwritten_extents(struct inode *inode,
+ u64 start, u64 len)
+{
+ int ret;
+ u32 cpos, phys_cpos, clusters, alloc_size;
+ u64 end = start + len;
+ struct buffer_head *di_bh = NULL;
+
+ if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
+ ret = ocfs2_read_inode_block(inode, &di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ /*
+ * Nothing to do if the requested reservation range
+ * fits within the inode.
+ */
+ if (ocfs2_size_fits_inline_data(di_bh, end))
+ goto out;
+
+ ret = ocfs2_convert_inline_data_to_extents(inode, di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ /*
+ * We consider both start and len to be inclusive.
+ */
+ cpos = start >> OCFS2_SB(inode->i_sb)->s_clustersize_bits;
+ clusters = ocfs2_clusters_for_bytes(inode->i_sb, start + len);
+ clusters -= cpos;
+
+ while (clusters) {
+ ret = ocfs2_get_clusters(inode, cpos, &phys_cpos,
+ &alloc_size, NULL);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ /*
+ * Hole or existing extent len can be arbitrary, so
+ * cap it to our own allocation request.
+ */
+ if (alloc_size > clusters)
+ alloc_size = clusters;
+
+ if (phys_cpos) {
+ /*
+ * We already have an allocation at this
+ * region so we can safely skip it.
+ */
+ goto next;
+ }
+
+ ret = __ocfs2_extend_allocation(inode, cpos, alloc_size, 1);
+ if (ret) {
+ if (ret != -ENOSPC)
+ mlog_errno(ret);
+ goto out;
+ }
+
+next:
+ cpos += alloc_size;
+ clusters -= alloc_size;
+ }
+
+ ret = 0;
+out:
+
+ brelse(di_bh);
+ return ret;
+}
+
+/*
+ * Truncate a byte range, avoiding pages within partial clusters. This
+ * preserves those pages for the zeroing code to write to.
+ */
+static void ocfs2_truncate_cluster_pages(struct inode *inode, u64 byte_start,
+ u64 byte_len)
+{
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ loff_t start, end;
+ struct address_space *mapping = inode->i_mapping;
+
+ start = (loff_t)ocfs2_align_bytes_to_clusters(inode->i_sb, byte_start);
+ end = byte_start + byte_len;
+ end = end & ~(osb->s_clustersize - 1);
+
+ if (start < end) {
+ unmap_mapping_range(mapping, start, end - start, 0);
+ truncate_inode_pages_range(mapping, start, end - 1);
+ }
+}
+
+static int ocfs2_zero_partial_clusters(struct inode *inode,
+ u64 start, u64 len)
+{
+ int ret = 0;
+ u64 tmpend, end = start + len;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ unsigned int csize = osb->s_clustersize;
+ handle_t *handle;
+
+ /*
+ * The "start" and "end" values are NOT necessarily part of
+ * the range whose allocation is being deleted. Rather, this
+ * is what the user passed in with the request. We must zero
+ * partial clusters here. There's no need to worry about
+ * physical allocation - the zeroing code knows to skip holes.
+ */
+ mlog(0, "byte start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ /*
+ * If both edges are on a cluster boundary then there's no
+ * zeroing required as the region is part of the allocation to
+ * be truncated.
+ */
+ if ((start & (csize - 1)) == 0 && (end & (csize - 1)) == 0)
+ goto out;
+
+ handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ mlog_errno(ret);
+ goto out;
+ }
+
+ /*
+ * We want to get the byte offset of the end of the 1st cluster.
+ */
+ tmpend = (u64)osb->s_clustersize + (start & ~(osb->s_clustersize - 1));
+ if (tmpend > end)
+ tmpend = end;
+
+ mlog(0, "1st range: start: %llu, tmpend: %llu\n",
+ (unsigned long long)start, (unsigned long long)tmpend);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, tmpend);
+ if (ret)
+ mlog_errno(ret);
+
+ if (tmpend < end) {
+ /*
+ * This may make start and end equal, but the zeroing
+ * code will skip any work in that case so there's no
+ * need to catch it up here.
+ */
+ start = end & ~(osb->s_clustersize - 1);
+
+ mlog(0, "2nd range: start: %llu, end: %llu\n",
+ (unsigned long long)start, (unsigned long long)end);
+
+ ret = ocfs2_zero_range_for_truncate(inode, handle, start, end);
+ if (ret)
+ mlog_errno(ret);
+ }
+
+ ocfs2_commit_trans(osb, handle);
+out:
+ return ret;
+}
+
+static int ocfs2_remove_inode_range(struct inode *inode,
+ struct buffer_head *di_bh, u64 byte_start,
+ u64 byte_len)
+{
+ int ret = 0;
+ u32 trunc_start, trunc_len, cpos, phys_cpos, alloc_size;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct ocfs2_cached_dealloc_ctxt dealloc;
+ struct address_space *mapping = inode->i_mapping;
+ struct ocfs2_extent_tree et;
+
+ ocfs2_init_dinode_extent_tree(&et, inode, di_bh);
+ ocfs2_init_dealloc_ctxt(&dealloc);
+
+ if (byte_len == 0)
+ return 0;
+
+ if (OCFS2_I(inode)->ip_dyn_features & OCFS2_INLINE_DATA_FL) {
+ ret = ocfs2_truncate_inline(inode, di_bh, byte_start,
+ byte_start + byte_len, 0);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+ /*
+ * There's no need to get fancy with the page cache
+ * truncate of an inline-data inode. We're talking
+ * about less than a page here, which will be cached
+ * in the dinode buffer anyway.
+ */
+ unmap_mapping_range(mapping, 0, 0, 0);
+ truncate_inode_pages(mapping, 0);
+ goto out;
+ }
+
+ trunc_start = ocfs2_clusters_for_bytes(osb->sb, byte_start);
+ trunc_len = (byte_start + byte_len) >> osb->s_clustersize_bits;
+ if (trunc_len >= trunc_start)
+ trunc_len -= trunc_start;
+ else
+ trunc_len = 0;
+
+ mlog(0, "Inode: %llu, start: %llu, len: %llu, cstart: %u, clen: %u\n",
+ (unsigned long long)OCFS2_I(inode)->ip_blkno,
+ (unsigned long long)byte_start,
+ (unsigned long long)byte_len, trunc_start, trunc_len);
+
+ ret = ocfs2_zero_partial_clusters(inode, byte_start, byte_len);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ cpos = trunc_start;
+ while (trunc_len) {
+ ret = ocfs2_get_clusters(inode, cpos, &phys_cpos,
+ &alloc_size, NULL);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ if (alloc_size > trunc_len)
+ alloc_size = trunc_len;
+
+ /* Only do work for non-holes */
+ if (phys_cpos != 0) {
+ ret = ocfs2_remove_btree_range(inode, &et, cpos,
+ phys_cpos, alloc_size,
+ &dealloc);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+ }
+
+ cpos += alloc_size;
+ trunc_len -= alloc_size;
+ }
+
+ ocfs2_truncate_cluster_pages(inode, byte_start, byte_len);
+
+out:
+ ocfs2_schedule_truncate_log_flush(osb, 1);
+ ocfs2_run_deallocs(osb, &dealloc);
+
+ return ret;
+}
+
+/*
+ * Parts of this function taken from xfs_change_file_space()
+ */
+static int __ocfs2_change_file_space(struct file *file, struct inode *inode,
+ loff_t f_pos, unsigned int cmd,
+ struct ocfs2_space_resv *sr,
+ int change_size)
+{
+ int ret;
+ s64 llen;
+ loff_t size;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct buffer_head *di_bh = NULL;
+ handle_t *handle;
+ unsigned long long max_off = inode->i_sb->s_maxbytes;
+
+ if (ocfs2_is_hard_readonly(osb) || ocfs2_is_soft_readonly(osb))
+ return -EROFS;
+
+ mutex_lock(&inode->i_mutex);
+
+ /*
+ * This prevents concurrent writes on other nodes
+ */
+ ret = ocfs2_rw_lock(inode, 1);
+ if (ret) {
+ mlog_errno(ret);
+ goto out;
+ }
+
+ ret = ocfs2_inode_lock(inode, &di_bh, 1);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_rw_unlock;
+ }
+
+ if (inode->i_flags & (S_IMMUTABLE|S_APPEND)) {
+ ret = -EPERM;
+ goto out_inode_unlock;
+ }
+
+ switch (sr->l_whence) {
+ case 0: /*SEEK_SET*/
+ break;
+ case 1: /*SEEK_CUR*/
+ sr->l_start += f_pos;
+ break;
+ case 2: /*SEEK_END*/
+ sr->l_start += i_size_read(inode);
+ break;
+ default:
+ ret = -EINVAL;
+ goto out_inode_unlock;
+ }
+ sr->l_whence = 0;
+
+ llen = sr->l_len > 0 ? sr->l_len - 1 : sr->l_len;
+
+ if (sr->l_start < 0
+ || sr->l_start > max_off
+ || (sr->l_start + llen) < 0
+ || (sr->l_start + llen) > max_off) {
+ ret = -EINVAL;
+ goto out_inode_unlock;
+ }
+ size = sr->l_start + sr->l_len;
+
+ if (cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) {
+ if (sr->l_len <= 0) {
+ ret = -EINVAL;
+ goto out_inode_unlock;
+ }
+ }
+
+ if (file && should_remove_suid(file->f_path.dentry)) {
+ ret = __ocfs2_write_remove_suid(inode, di_bh);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_inode_unlock;
+ }
+ }
+
+ down_write(&OCFS2_I(inode)->ip_alloc_sem);
+ switch (cmd) {
+ case OCFS2_IOC_RESVSP:
+ case OCFS2_IOC_RESVSP64:
+ /*
+ * This takes unsigned offsets, but the signed ones we
+ * pass have been checked against overflow above.
+ */
+ ret = ocfs2_allocate_unwritten_extents(inode, sr->l_start,
+ sr->l_len);
+ break;
+ case OCFS2_IOC_UNRESVSP:
+ case OCFS2_IOC_UNRESVSP64:
+ ret = ocfs2_remove_inode_range(inode, di_bh, sr->l_start,
+ sr->l_len);
+ break;
+ default:
+ ret = -EINVAL;
+ }
+ up_write(&OCFS2_I(inode)->ip_alloc_sem);
+ if (ret) {
+ mlog_errno(ret);
+ goto out_inode_unlock;
+ }
+
+ /*
+ * We update c/mtime for these changes
+ */
+ handle = ocfs2_start_trans(osb, OCFS2_INODE_UPDATE_CREDITS);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ mlog_errno(ret);
+ goto out_inode_unlock;
+ }
+
+ if (change_size && i_size_read(inode) < size)
+ i_size_write(inode, size);
+
+ inode->i_ctime = inode->i_mtime = CURRENT_TIME;
+ ret = ocfs2_mark_inode_dirty(handle, inode, di_bh);
+ if (ret < 0)
+ mlog_errno(ret);
+
+ ocfs2_commit_trans(osb, handle);
+
+out_inode_unlock:
+ brelse(di_bh);
+ ocfs2_inode_unlock(inode, 1);
+out_rw_unlock:
+ ocfs2_rw_unlock(inode, 1);
+
+out:
+ mutex_unlock(&inode->i_mutex);
+ return ret;
+}
+
+int ocfs2_change_file_space(struct file *file, unsigned int cmd,
+ struct ocfs2_space_resv *sr)
+{
+ struct inode *inode = file->f_path.dentry->d_inode;
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+
+ if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) &&
+ !ocfs2_writes_unwritten_extents(osb))
+ return -ENOTTY;
+ else if ((cmd == OCFS2_IOC_UNRESVSP || cmd == OCFS2_IOC_UNRESVSP64) &&
+ !ocfs2_sparse_alloc(osb))
+ return -ENOTTY;
+
+ if (!S_ISREG(inode->i_mode))
+ return -EINVAL;
+
+ if (!(file->f_mode & FMODE_WRITE))
+ return -EBADF;
+
+ return __ocfs2_change_file_space(file, inode, file->f_pos, cmd, sr, 0);
+}
+
+static long ocfs2_fallocate(struct inode *inode, int mode, loff_t offset,
+ loff_t len)
+{
+ struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
+ struct ocfs2_space_resv sr;
+ int change_size = 1;
+
+ if (!ocfs2_writes_unwritten_extents(osb))
+ return -EOPNOTSUPP;
+
+ if (S_ISDIR(inode->i_mode))
+ return -ENODEV;
+
+ if (mode & FALLOC_FL_KEEP_SIZE)
+ change_size = 0;
+
+ sr.l_whence = 0;
+ sr.l_start = (s64)offset;
+ sr.l_len = (s64)len;
+
+ return __ocfs2_change_file_space(NULL, inode, offset,
+ OCFS2_IOC_RESVSP64, &sr, change_size);
+}
+