Linux Kernel文件系统写I/O流程代码分析（二）bdi_writeback

Linux Kernel文件系统写I/O流程代码分析（二）bdi_writeback

上一篇# Linux Kernel文件系统写I/O流程代码分析（一），我们看到Buffered IO，写操作写入到page cache后就直接返回了，本文主要分析脏页是如何刷盘的。

概述

由于内核page cache的作用，写操作实际被延迟写入。当page cache里的数据被用户写入但是没有刷新到磁盘时，则该page为脏页（块设备page cache机制因为以前机械磁盘以扇区为单位读写，引入了buffer_head，每个4K的page进一步划分成8个buffer，通过buffer_head管理，因此可能只设置了部分buffer head为脏）。

脏页在以下情况下将被回写（write back）到磁盘上：

• 脏页在内存里的时间超过了阈值。
• 系统的内存紧张，低于某个阈值时，必须将所有脏页回写。
• 用户强制要求刷盘，如调用sync()、fsync()、close()等系统调用。

以前的Linux通过pbflush机制管理脏页的回写，但因为其管理了所有的磁盘的page/buffer_head，存在严重的性能瓶颈，因此从Linux 2.6.32开始，脏页回写的工作由bdi_writeback机制负责。bdi_writeback机制为每个磁盘都创建一个线程，专门负责这个磁盘的page cache或者

buffer cache的数据刷新工作，以提高I/O性能。

BDI系统

BDI是backing device info的缩写，它用于描述后端存储（如磁盘）设备相关的信息。相对于内存来说，后端存储的I/O比较慢，因此写盘操作需要通过page cache进行缓存延迟写入。

最初的BDI子系统里，模块启动的时候创建bdi-default进程，然后为每个注册的设备创建flush-x:y（x,y为主次设备号）的进程，用于脏数据的回写。在Linux 3.10.0版本之后，BDI子系统使用workqueue机制代替原来的线程创建，需要回写时，将flush任务提交给workqueue，最终由通用的[kworker]进程负责处理。BDI子系统初始化的代码如下：

static int __init default_bdi_init(void)
{
	int err;

	bdi_wq = alloc_workqueue("writeback", WQ_MEM_RECLAIM | WQ_FREEZABLE |
					      WQ_UNBOUND | WQ_SYSFS, 0);
	if (!bdi_wq)
		return -ENOMEM;

	err = bdi_init(&default_backing_dev_info);
	if (!err)
		bdi_register(&default_backing_dev_info, NULL, "default");
	err = bdi_init(&noop_backing_dev_info);

	return err;
}
subsys_initcall(default_bdi_init);

设备注册

当执行mount流程时，底层文件系统定义自己的struct backing_dev_info结构并将其注册到BDI子系统，如下是FUSE代码示例：

static int fuse_bdi_init(struct fuse_conn *fc, struct super_block *sb)
{
	int err;

	fc->bdi.name = "fuse";
	fc->bdi.ra_pages = (VM_MAX_READAHEAD * 1024) / PAGE_CACHE_SIZE;
	/* fuse does it's own writeback accounting */
	fc->bdi.capabilities = BDI_CAP_NO_ACCT_WB | BDI_CAP_STRICTLIMIT;

	err = bdi_init(&fc->bdi);
	if (err)
		return err;

	fc->bdi_initialized = 1;

	if (sb->s_bdev) {
		err =  bdi_register(&fc->bdi, NULL, "%u:%u-fuseblk",
				    MAJOR(fc->dev), MINOR(fc->dev));
	} else {
		err = bdi_register_dev(&fc->bdi, fc->dev);
	}

	if (err)
		return err;

	/*
	 *    /sys/class/bdi/<bdi>/max_ratio
	 */
	bdi_set_max_ratio(&fc->bdi, 1);

	return 0;
}

该函数先通过bdi_init()初始化struct backing_dev_info，然后通过bid_register()将其注册到BDI子系统。

其中bdi_init()会调用bdi_wb_init()初始化struct bdi_writeback：

static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
{
	memset(wb, 0, sizeof(*wb));

	wb->bdi = bdi;
	wb->last_old_flush = jiffies;
	INIT_LIST_HEAD(&wb->b_dirty);
	INIT_LIST_HEAD(&wb->b_io);
	INIT_LIST_HEAD(&wb->b_more_io);
	spin_lock_init(&wb->list_lock);
	INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
}

其中初始化了一个默认处理函数为bdi_writeback_workfn的work，用于回写处理。

数据回写

在上一篇的基础上，将图补充了bdi回写的部分，如下所示：

bdi_queue_work

BDI子系统使用workqueue机制进行数据回写，其回写接口为bdi_queue_work()将具体某个bdi的回写请求（wb_writeback_work）挂到bdi_wq上。代码如下：

static void bdi_queue_work(struct backing_dev_info *bdi,
			   struct wb_writeback_work *work)
{
	trace_writeback_queue(bdi, work);

	spin_lock_bh(&bdi->wb_lock);
	if (!test_bit(BDI_registered, &bdi->state)) {
		if (work->done)
			complete(work->done);
		goto out_unlock;
	}
	list_add_tail(&work->list, &bdi->work_list);
	mod_delayed_work(bdi_wq, &bdi->wb.dwork, 0);
out_unlock:
	spin_unlock_bh(&bdi->wb_lock);
}

调用该函数的地方包括：

• sync_inode_sb(): 将该super block上所有的脏inode回写。
• writeback_inodes_sb_nr():回写super block上指定个数脏inode。
• __bdi_start_writeback():定时调用或者需要释放pages或者需要更多内存时调用。

bdi_writeback_workfn

bdi_queue_work()提交了work给bdi_wq上，由对应的bdi处理函数进行处理，默认的函数为bdi_writeback_workfn，其代码如下：

void bdi_writeback_workfn(struct work_struct *work)
{
	struct bdi_writeback *wb = container_of(to_delayed_work(work),
						struct bdi_writeback, dwork);
	struct backing_dev_info *bdi = wb->bdi;
	long pages_written;

	set_worker_desc("flush-%s", dev_name(bdi->dev));
	current->flags |= PF_SWAPWRITE;

	if (likely(!current_is_workqueue_rescuer() ||
		   !test_bit(BDI_registered, &bdi->state))) {
		/*
		 * The normal path.  Keep writing back @bdi until its
		 * work_list is empty.  Note that this path is also taken
		 * if @bdi is shutting down even when we're running off the
		 * rescuer as work_list needs to be drained.
		 */
		do {
			pages_written = wb_do_writeback(wb);
			trace_writeback_pages_written(pages_written);
		} while (!list_empty(&bdi->work_list));
	} else {
		/*
		 * bdi_wq can't get enough workers and we're running off
		 * the emergency worker.  Don't hog it.  Hopefully, 1024 is
		 * enough for efficient IO.
		 */
		pages_written = writeback_inodes_wb(&bdi->wb, 1024,
						    WB_REASON_FORKER_THREAD);
		trace_writeback_pages_written(pages_written);
	}

	if (!list_empty(&bdi->work_list))
		mod_delayed_work(bdi_wq, &wb->dwork, 0);
	else if (wb_has_dirty_io(wb) && dirty_writeback_interval)
		bdi_wakeup_thread_delayed(bdi);

	current->flags &= ~PF_SWAPWRITE;
}

首先判断当前workqueue能否获得足够的worker进行处理，如果能则将bdi上所有work全部提交，否则只提交一个work并限制写入1024个pages。

正常情况下通过调用wb_do_writeback函数处理回写。

wb_do_writeback

该函数代码如下，遍历bdi上所有work，通过调用wb_writeback()进行数据写入。

static long wb_do_writeback(struct bdi_writeback *wb)
{
	struct backing_dev_info *bdi = wb->bdi;
	struct wb_writeback_work *work;
	long wrote = 0;

	set_bit(BDI_writeback_running, &wb->bdi->state);
	while ((work = get_next_work_item(bdi)) != NULL) {

		trace_writeback_exec(bdi, work);

		wrote += wb_writeback(wb, work);

		/*
		 * Notify the caller of completion if this is a synchronous
		 * work item, otherwise just free it.
		 */
		if (work->done)
			complete(work->done);
		else
			kfree(work);
	}

	/*
	 * Check for periodic writeback, kupdated() style
	 */
	wrote += wb_check_old_data_flush(wb);
	wrote += wb_check_background_flush(wb);
	clear_bit(BDI_writeback_running, &wb->bdi->state);

	return wrote;
}

wb_writeback()函数最终调用__writeback_single_inode()将某个inode上脏页刷回。

__writeback_single_inode

__writeback_single_inode()的代码如下，最终通过调用do_writepages()函数写盘：

static int
__writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
{
	struct address_space *mapping = inode->i_mapping;
	long nr_to_write = wbc->nr_to_write;
	unsigned dirty;
	int ret;

	WARN_ON(!(inode->i_state & I_SYNC));

	trace_writeback_single_inode_start(inode, wbc, nr_to_write);

	ret = do_writepages(mapping, wbc);

	/*
	 * Make sure to wait on the data before writing out the metadata.
	 * This is important for filesystems that modify metadata on data
	 * I/O completion. We don't do it for sync(2) writeback because it has a
	 * separate, external IO completion path and ->sync_fs for guaranteeing
	 * inode metadata is written back correctly.
	 */
	if (wbc->sync_mode == WB_SYNC_ALL && !wbc->for_sync) {
		int err = filemap_fdatawait(mapping);
		if (ret == 0)
			ret = err;
	}

	/*
	 * Some filesystems may redirty the inode during the writeback
	 * due to delalloc, clear dirty metadata flags right before
	 * write_inode()
	 */
	spin_lock(&inode->i_lock);
	/* Clear I_DIRTY_PAGES if we've written out all dirty pages */
	if (!mapping_tagged(mapping, PAGECACHE_TAG_DIRTY))
		inode->i_state &= ~I_DIRTY_PAGES;
	dirty = inode->i_state & I_DIRTY;
	inode->i_state &= ~(I_DIRTY_SYNC | I_DIRTY_DATASYNC);
	spin_unlock(&inode->i_lock);
	/* Don't write the inode if only I_DIRTY_PAGES was set */
	if (dirty & (I_DIRTY_SYNC | I_DIRTY_DATASYNC)) {
		int err = write_inode(inode, wbc);
		if (ret == 0)
			ret = err;
	}
	trace_writeback_single_inode(inode, wbc, nr_to_write);
	return ret;
}

do_writepages

函数do_writepages()在上一篇已经介绍过了，它负责调用底层文件系统的a_ops->writepages将pages写入后端存储。