[Ocfs2-devel] [PATCH 04/39] ocfs2: Basic tree root operation.
Joel Becker
Joel.Becker at oracle.com
Fri May 1 20:55:55 PDT 2009
On Thu, Apr 30, 2009 at 06:58:16AM +0800, Tao Ma wrote:
> +int ocfs2_create_refcount_tree(struct inode *inode, struct buffer_head *di_bh)
> +{
> + int ret;
> + handle_t *handle = NULL;
> + struct ocfs2_alloc_context *meta_ac = NULL;
> + struct ocfs2_dinode *di = (struct ocfs2_dinode *)di_bh->b_data;
> + struct ocfs2_inode_info *oi = OCFS2_I(inode);
> + struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
> + struct buffer_head *new_bh = NULL;
> + struct ocfs2_refcount_block *rb;
> + u16 suballoc_bit_start;
> + u32 num_got;
> + u64 first_blkno;
> +
> + BUG_ON(oi->ip_dyn_features & OCFS2_HAS_REFCOUNT_FL);
> + BUG_ON(di->i_refcount_loc);
> +
> + mlog(0, "create tree for inode %lu\n", inode->i_ino);
> +
> + ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
> + if (ret) {
> + mlog_errno(ret);
> + goto out;
> + }
> +
> + handle = ocfs2_start_trans(osb, OCFS2_REFCOUNT_TREE_CREATE_CREDITS);
> + if (IS_ERR(handle)) {
> + ret = PTR_ERR(handle);
> + mlog_errno(ret);
> + goto out;
> + }
> +
> + ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode), di_bh,
> + OCFS2_JOURNAL_ACCESS_WRITE);
> + if (ret) {
> + mlog_errno(ret);
> + goto out_commit;
> + }
> +
> + ret = ocfs2_claim_metadata(osb, handle, meta_ac, 1,
> + &suballoc_bit_start, &num_got,
> + &first_blkno);
> + if (ret) {
> + mlog_errno(ret);
> + goto out_commit;
> + }
> +
> + new_bh = sb_getblk(inode->i_sb, first_blkno);
> + ocfs2_set_new_buffer_uptodate(INODE_CACHE(inode), new_bh);
> +
> + ret = ocfs2_journal_access_rb(handle, INODE_CACHE(inode), new_bh,
> + OCFS2_JOURNAL_ACCESS_CREATE);
> + if (ret) {
> + mlog_errno(ret);
> + goto out_commit;
> + }
> +
> + /* Initialize ocfs2_refcount_block. */
> + rb = (struct ocfs2_refcount_block *)new_bh->b_data;
> + memset(rb, 0, inode->i_sb->s_blocksize);
> + strcpy((void *)rb, OCFS2_REFCOUNT_BLOCK_SIGNATURE);
> + rb->rf_suballoc_slot = cpu_to_le16(osb->slot_num);
> + rb->rf_suballoc_bit = cpu_to_le16(suballoc_bit_start);
> + rb->rf_fs_generation = cpu_to_le32(osb->fs_generation);
> + rb->rf_blkno = cpu_to_le64(first_blkno);
> + rb->rf_count = cpu_to_le32(1);
> + rb->rf_flags = cpu_to_le32(OCFS2_REFCOUNT_LEAF_FL);
> + rb->rf_records.rl_count =
> + cpu_to_le16(ocfs2_refcount_recs_per_rb(osb->sb));
Ouch, I missed this the first pass. The root of the tree is not
a LEAF, so it should have flags==0. OCFS2_REFCOUNT_LEAF_FL does not
mean "I have refcount records". It means "I am hanging off of a btree
and rf_parent points to the top of my btree". To determine whether
rf_records is valid, you check OCFS2_REFCOUNT_TREE_FL. If not set,
rf_records is valid.
Joel
--
"Always give your best, never get discouraged, never be petty; always
remember, others may hate you. Those who hate you don't win unless
you hate them. And then you destroy yourself."
- Richard M. Nixon
Joel Becker
Principal Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
More information about the Ocfs2-devel
mailing list