标签：
在做进程安全监控的时候，拍脑袋决定的，如果发现一个进程在D状态时，即TASK_UNINTERRUPTIBLE（不可中断的睡眠状态），时间超过了8min，就将系统panic掉。恰好DB组做日志时，将整个log缓存到内存中，最后刷磁盘，结果系统就D状态了很长时间，自然panic了，中间涉及到Linux的缓存写回刷磁盘的一些机制和调优方法，写一下总结。
目前机制需要将脏页刷回到磁盘一般是以下情况：
内核使用pdflush线程刷新脏页到磁盘，pdflush线程个数在2和8之间，可以通过/proc/sys/vm/nr_pdflush_threads文件直接查看，具体策略机制参看源码函数__pdflush。
先说一下第一种和第三种情况：当内存空间不足或外界强制刷新的时候，脏页的刷新是通过调用wakeup_pdflush函数实现的，调用其函数的有do_sync、free_more_memory、try_to_free_pages。wakeup_pdflush的功能是通过background_writeout的函数实现的：
static void background_writeout(unsigned long _min_pages) { long min_pages = _min_pages; struct writeback_control wbc = { .bdi = NULL, .sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .nr_to_write = 0, .nonblocking = 1, }; for ( ; ; ) { struct writeback_state wbs; long background_thresh; long dirty_thresh; get_dirty_limits(&wbs, &background_thresh, &dirty_thresh, NULL); if (wbs.nr_dirty + wbs.nr_unstable < background_thresh && min_pages <= 0) break; wbc.encountered_congestion = 0; wbc.nr_to_write = MAX_WRITEBACK_PAGES; wbc.pages_skipped = 0; writeback_inodes(&wbc); min_pages -= MAX_WRITEBACK_PAGES - wbc.nr_to_write; if (wbc.nr_to_write > 0 || wbc.pages_skipped > 0) { /* Wrote less than expected */ blk_congestion_wait(WRITE, HZ/10); if (!wbc.encountered_congestion) break; } } }
background_writeout进到一个死循环里面，通过get_dirty_limits获取脏页开始刷新的临界值background_thresh，即为dirty_background_ratio的总内存页数百分比，可以通过proc接口/proc/sys/vm/dirty_background_ratio调整，一般默认为10。当脏页超过临界值时，调用writeback_inodes写MAX_WRITEBACK_PAGES（1024）个页，直到脏页比例低于临界值。
内核在启动的时候在page_writeback_init初始化wb_timer定时器，超时时间是dirty_writeback_centisecs，单位是0.01秒，可以通过/proc/sys/vm/dirty_writeback_centisecs调节。wb_timer的触发函数是wb_timer_fn，最终是通过wb_kupdate实现。
static void wb_kupdate(unsigned long arg) { sync_supers(); get_writeback_state(&wbs); oldest_jif = jiffies - (dirty_expire_centisecs * HZ) / 100; start_jif = jiffies; next_jif = start_jif + (dirty_writeback_centisecs * HZ) / 100; nr_to_write = wbs.nr_dirty + wbs.nr_unstable + (inodes_stat.nr_inodes - inodes_stat.nr_unused); while (nr_to_write > 0) { wbc.encountered_congestion = 0; wbc.nr_to_write = MAX_WRITEBACK_PAGES; writeback_inodes(&wbc); if (wbc.nr_to_write > 0) { if (wbc.encountered_congestion) blk_congestion_wait(WRITE, HZ/10); else break; /* All the old data is written */ } nr_to_write -= MAX_WRITEBACK_PAGES - wbc.nr_to_write; } if (time_before(next_jif, jiffies + HZ)) next_jif = jiffies + HZ; if (dirty_writeback_centisecs) mod_timer(&wb_timer, next_jif); }
上面的代码没有拷贝全。内核首先将超级块信息刷新到文件系统上，然后获取oldest_jif作为wbc的参数只刷新已修改时间大于dirty_expire_centisecs的脏页，dirty_expire_centisecs参数可以通过/proc/sys/vm/dirty_expire_centisecs调整。
用户态使用WRITE函数写文件时也有可能要刷新脏页，generic_file_buffered_write函数会在将写的内存页标记为脏之后，根据条件刷新磁盘以平衡当前脏页比率，参看balance_dirty_pages_ratelimited函数：
void balance_dirty_pages_ratelimited(struct address_space *mapping) { static DEFINE_PER_CPU(int, ratelimits) = 0; long ratelimit; ratelimit = ratelimit_pages; if (dirty_exceeded) ratelimit = 8; /* * Check the rate limiting. Also, we do not want to throttle real-time * tasks in balance_dirty_pages(). Period. */ if (get_cpu_var(ratelimits)++ >= ratelimit) { __get_cpu_var(ratelimits) = 0; put_cpu_var(ratelimits); balance_dirty_pages(mapping); return; } put_cpu_var(ratelimits); }
balance_dirty_pages_ratelimited函数通过ratelimit_pages调节刷新（调用balance_dirty_pages函数）的次数，每ratelimit_pages次调用才会刷新一次，具体刷新过程看balance_dirty_pages函数：
static void balance_dirty_pages(struct address_space *mapping) { struct writeback_state wbs; long nr_reclaimable; long background_thresh; long dirty_thresh; unsigned long pages_written = 0; unsigned long write_chunk = sync_writeback_pages(); struct backing_dev_info *bdi = mapping->backing_dev_info; for (;;) { struct writeback_control wbc = { .bdi = bdi, .sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .nr_to_write = write_chunk, }; get_dirty_limits(&wbs, &background_thresh, &dirty_thresh, mapping); nr_reclaimable = wbs.nr_dirty + wbs.nr_unstable; if (nr_reclaimable + wbs.nr_writeback <= dirty_thresh) break; if (!dirty_exceeded) dirty_exceeded = 1; /* Note: nr_reclaimable denotes nr_dirty + nr_unstable. * Unstable writes are a feature of certain networked * filesystems (i.e. NFS) in which data may have been * written to the server‘s write cache, but has not yet * been flushed to permanent storage. */ if (nr_reclaimable) { writeback_inodes(&wbc); get_dirty_limits(&wbs, &background_thresh, &dirty_thresh, mapping); nr_reclaimable = wbs.nr_dirty + wbs.nr_unstable; if (nr_reclaimable + wbs.nr_writeback <= dirty_thresh) break; pages_written += write_chunk - wbc.nr_to_write; if (pages_written >= write_chunk) break; /* We‘ve done our duty */ } blk_congestion_wait(WRITE, HZ/10); } if (nr_reclaimable + wbs.nr_writeback <= dirty_thresh && dirty_exceeded) dirty_exceeded = 0; if (writeback_in_progress(bdi)) return; /* pdflush is already working this queue */ /* * In laptop mode, we wait until hitting the higher threshold before * starting background writeout, and then write out all the way down * to the lower threshold. So slow writers cause minimal disk activity. * * In normal mode, we start background writeout at the lower * background_thresh, to keep the amount of dirty memory low. */ if ((laptop_mode && pages_written) || (!laptop_mode && (nr_reclaimable > background_thresh))) pdflush_operation(background_writeout, 0); }
函数走进一个死循环，通过get_dirty_limits获取dirty_background_ratio和dirty_ratio对应的内存页数值，当24行做判断，如果脏页大于dirty_thresh，则调用writeback_inodes开始刷缓存到磁盘，如果一次没有将脏页比率刷到dirty_ratio之下，则用blk_congestion_wait阻塞写，然后反复循环，直到比率降低到dirty_ratio；当比率低于dirty_ratio之后，但脏页比率大于dirty_background_ratio，则用pdflush_operation启用background_writeout，pdflush_operation是非阻塞函数，唤醒pdflush后直接返回，background_writeout在有pdflush调用。
如此可知：WRITE写的时候，缓存超过dirty_ratio，则会阻塞写操作，回刷脏页，直到缓存低于dirty_ratio；如果缓存高于background_writeout，则会在写操作时，唤醒pdflush进程刷脏页，不阻塞写操作。
导致进程D状态大部分是因为第3种和第4种情况：有大量写操作，缓存由Linux系统管理，一旦脏页累计到一定程度，无论是继续写还是fsync刷新，都会使进程D住。
标签：
原文地址：http://www.cnblogs.com/wuchanming/p/4820830.html
