writeback: bdi write bandwidth estimation

The estimation value will start from 100MB/s and adapt to the real
bandwidth in seconds.

It tries to update the bandwidth only when disk is fully utilized.
Any inactive period of more than one second will be skipped.

The estimated bandwidth will be reflecting how fast the device can
writeout when _fully utilized_, and won't drop to 0 when it goes idle.
The value will remain constant at disk idle time. At busy write time, if
not considering fluctuations, it will also remain high unless be knocked
down by possible concurrent reads that compete for the disk time and
bandwidth with async writes.

The estimation is not done purely in the flusher because there is no
guarantee for write_cache_pages() to return timely to update bandwidth.

The bdi->avg_write_bandwidth smoothing is very effective for filtering
out sudden spikes, however may be a little biased in long term.

The overheads are low because the bdi bandwidth update only occurs at
200ms intervals.

The 200ms update interval is suitable, because it's not possible to get
the real bandwidth for the instance at all, due to large fluctuations.

The NFS commits can be as large as seconds worth of data. One XFS
completion may be as large as half second worth of data if we are going
to increase the write chunk to half second worth of data. In ext4,
fluctuations with time period of around 5 seconds is observed. And there
is another pattern of irregular periods of up to 20 seconds on SSD tests.

That's why we are not only doing the estimation at 200ms intervals, but
also averaging them over a period of 3 seconds and then go further to do
another level of smoothing in avg_write_bandwidth.

CC: Li Shaohua <shaohua.li@intel.com>
CC: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>

2 files changed, 99 insertions, 0 deletions

diff --git a/mm/backing-dev.c b/mm/backing-dev.c
index 83f18a1d9d10..a76cdd160277 100644
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -638,6 +638,11 @@ static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
 	setup_timer(&wb->wakeup_timer, wakeup_timer_fn, (unsigned long)bdi);
 }
 
+/*
+ * Initial write bandwidth: 100 MB/s
+ */
+#define INIT_BW		(100 << (20 - PAGE_SHIFT))
+
 int bdi_init(struct backing_dev_info *bdi)
 {
 	int i, err;
@@ -660,6 +665,13 @@ int bdi_init(struct backing_dev_info *bdi)
 	}
 
 	bdi->dirty_exceeded = 0;
+
+	bdi->bw_time_stamp = jiffies;
+	bdi->written_stamp = 0;
+
+	bdi->write_bandwidth = INIT_BW;
+	bdi->avg_write_bandwidth = INIT_BW;
+
 	err = prop_local_init_percpu(&bdi->completions);
 
 	if (err) {
diff --git a/mm/page-writeback.c b/mm/page-writeback.c
index 8cd71376c63d..446bdf7b975b 100644
--- a/mm/page-writeback.c
+++ b/mm/page-writeback.c
@@ -37,6 +37,11 @@
 #include <trace/events/writeback.h>
 
 /*
+ * Estimate write bandwidth at 200ms intervals.
+ */
+#define BANDWIDTH_INTERVAL	max(HZ/5, 1)
+
+/*
  * After a CPU has dirtied this many pages, balance_dirty_pages_ratelimited
  * will look to see if it needs to force writeback or throttling.
  */
@@ -471,6 +476,85 @@ unsigned long bdi_dirty_limit(struct backing_dev_info *bdi, unsigned long dirty)
 	return bdi_dirty;
 }
 
+static void bdi_update_write_bandwidth(struct backing_dev_info *bdi,
+				       unsigned long elapsed,
+				       unsigned long written)
+{
+	const unsigned long period = roundup_pow_of_two(3 * HZ);
+	unsigned long avg = bdi->avg_write_bandwidth;
+	unsigned long old = bdi->write_bandwidth;
+	u64 bw;
+
+	/*
+	 * bw = written * HZ / elapsed
+	 *
+	 *                   bw * elapsed + write_bandwidth * (period - elapsed)
+	 * write_bandwidth = ---------------------------------------------------
+	 *                                          period
+	 */
+	bw = written - bdi->written_stamp;
+	bw *= HZ;
+	if (unlikely(elapsed > period)) {
+		do_div(bw, elapsed);
+		avg = bw;
+		goto out;
+	}
+	bw += (u64)bdi->write_bandwidth * (period - elapsed);
+	bw >>= ilog2(period);
+
+	/*
+	 * one more level of smoothing, for filtering out sudden spikes
+	 */
+	if (avg > old && old >= (unsigned long)bw)
+		avg -= (avg - old) >> 3;
+
+	if (avg < old && old <= (unsigned long)bw)
+		avg += (old - avg) >> 3;
+
+out:
+	bdi->write_bandwidth = bw;
+	bdi->avg_write_bandwidth = avg;
+}
+
+void __bdi_update_bandwidth(struct backing_dev_info *bdi,
+			    unsigned long start_time)
+{
+	unsigned long now = jiffies;
+	unsigned long elapsed = now - bdi->bw_time_stamp;
+	unsigned long written;
+
+	/*
+	 * rate-limit, only update once every 200ms.
+	 */
+	if (elapsed < BANDWIDTH_INTERVAL)
+		return;
+
+	written = percpu_counter_read(&bdi->bdi_stat[BDI_WRITTEN]);
+
+	/*
+	 * Skip quiet periods when disk bandwidth is under-utilized.
+	 * (at least 1s idle time between two flusher runs)
+	 */
+	if (elapsed > HZ && time_before(bdi->bw_time_stamp, start_time))
+		goto snapshot;
+
+	bdi_update_write_bandwidth(bdi, elapsed, written);
+
+snapshot:
+	bdi->written_stamp = written;
+	bdi->bw_time_stamp = now;
+}
+
+static void bdi_update_bandwidth(struct backing_dev_info *bdi,
+				 unsigned long start_time)
+{
+	if (time_is_after_eq_jiffies(bdi->bw_time_stamp + BANDWIDTH_INTERVAL))
+		return;
+	spin_lock(&bdi->wb.list_lock);
+	__bdi_update_bandwidth(bdi, start_time);
+	spin_unlock(&bdi->wb.list_lock);
+}
+
 /*
  * balance_dirty_pages() must be called by processes which are generating dirty
  * data.  It looks at the number of dirty pages in the machine and will force
@@ -490,6 +574,7 @@ static void balance_dirty_pages(struct address_space *mapping,
 	unsigned long pause = 1;
 	bool dirty_exceeded = false;
 	struct backing_dev_info *bdi = mapping->backing_dev_info;
+	unsigned long start_time = jiffies;
 
 	for (;;) {
 		nr_reclaimable = global_page_state(NR_FILE_DIRTY) +
@@ -544,6 +629,8 @@ static void balance_dirty_pages(struct address_space *mapping,
 		if (!bdi->dirty_exceeded)
 			bdi->dirty_exceeded = 1;
 
+		bdi_update_bandwidth(bdi, start_time);
+
 		/* 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