qemu commit实现简记

上一篇 http://blog.chinaunix.net/uid-29718549-id-4445694.htm 说的在线快照的一点问题，这次记录一下在线删除快照的核心流程，对应qemu 的 block-commit命令，就是实现在qcow2磁盘链表中删除磁盘，合并数据的过程

变量名对应的链表视图：
    -> [BASE] -> [INTER1] -> [INTER2] -> [TOP]

end = s->common.len >> BDRV_SECTOR_BITS;
    buf = qemu_blockalign(top, COMMIT_BUFFER_SIZE);
    for (sector_num = 0; sector_num < end; sector_num += n) {\\遍历top中所有的块
        uint64_t delay_ns = 0;
        bool copy;

wait:
        /* Note that even when no rate limit is applied we need to yield
         * with no pending I/O here so that bdrv_drain_all() returns.
         */
        block_job_sleep_ns(&s->common, QEMU_CLOCK_REALTIME, delay_ns);
        if (block_job_is_cancelled(&s->common)) {
            break;
        }
        /* Copy if allocated above the base */
        /*以下这一坨代码负责检查指定范围内（一次检查1024个sector）需不需要执行commit*/
        ret = bdrv_is_allocated_above(top, base, sector_num,
                                      COMMIT_BUFFER_SIZE / BDRV_SECTOR_SIZE, \\一次检查1024个块
                                      &n);\\检查这个范围的块是否在链表中任意磁盘中被写入过，如果有写入（返回值1）->执行以下的copy动作
            intermediate = top;
            while (intermediate && intermediate != base)\\遍历链表top(包含top)->base的所有中间部分磁盘
                ret = bdrv_is_allocated(intermediate, sector_num, nb_sectors,
                                &pnum_inter);
                    ret = bdrv_get_block_status(bs, sector_num, nb_sectors, pnum);
                        bdrv_get_block_status_co_entry
                            bdrv_co_get_block_status(bs, data->sector_num, data->nb_sectors,
                                         data->pnum);\\<qcow2_co_get_block_status>
                                *pnum = nb_sectors;
                                qemu_co_mutex_lock(&s->lock);\\获取状态码之前加锁
                                ret = qcow2_get_cluster_offset(bs, sector_num << 9, pnum, &cluster_offset);\\下面这坨下会详解
                                    l1_bits = s->l2_bits + s->cluster_bits;
                                    l1_index = offset >> l1_bits;\\这里的offset就是sector_num << 9，也就是seek点
                                    if (l1_index >= s->l1_size) ret = QCOW2_CLUSTER_UNALLOCATED;\\offset超出了BDRVQcowState记录的L1表范围，返回未分配
                                    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
                                    if (l2_offset == NULL) ret=QCOW2_CLUSTER_UNALLOCATED\\在L1 table中无记录，返回未分配
                                    ret = l2_load(bs, l2_offset, &l2_table);\\读出L2 table
                                    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
                                    nb_clusters = size_to_clusters(s, nb_needed << 9);
                                    ret = qcow2_get_cluster_type(*cluster_offset);
                                qemu_co_mutex_unlock(&s->lock);\\获取状态码之后解锁
        /*以上这一坨代码负责检查指定范围内（一次检查1024个sector）需不需要执行commit*/
        copy = (ret == 1);
        trace_commit_one_iteration(s, sector_num, n, ret);
        if (copy) {
            if (s->common.speed) {\\这部分限速控制
                delay_ns = ratelimit_calculate_delay(&s->limit, n);
                if (delay_ns > 0) {
                    goto wait;
                }
            }
            ret = commit_populate(top, base, sector_num, n, buf);\\读出->写入
                ret = bdrv_read(bs, sector_num, buf, nb_sectors);
                ret = bdrv_write(base, sector_num, buf, nb_sectors);
            bytes_written += n * BDRV_SECTOR_SIZE;
        }
        if (ret < 0) {
            if (s->on_error == BLOCKDEV_ON_ERROR_STOP ||
                s->on_error == BLOCKDEV_ON_ERROR_REPORT||
                (s->on_error == BLOCKDEV_ON_ERROR_ENOSPC && ret == -ENOSPC)) {
                goto exit_free_buf;
            } else {
                n = 0;
                continue;
            }
        }
        /* Publish progress */
        s->common.offset += n * BDRV_SECTOR_SIZE;
    }

    ret = 0;

    if (!block_job_is_cancelled(&s->common) && sector_num == end) {
        /* success */
        ret = bdrv_drop_intermediate(active, top, base);\\丢弃中间的部分，里面是改点属性
    }