linux-mm.kvack.org archive mirror
 help / color / mirror / Atom feed
From: Andrew Morton <akpm@linux-foundation.org>
To: Dave Chinner <david@fromorbit.com>
Cc: Jan Kara <jack@suse.cz>,
	linux-fsdevel@vger.kernel.org, linux-mm@kvack.org,
	hch@infradead.org, peterz@infradead.org, wfg@mail.ustc.edu.cn
Subject: Re: [PATCH RFC] mm: Implement balance_dirty_pages() through waiting for flusher thread
Date: Mon, 21 Jun 2010 23:14:16 -0700	[thread overview]
Message-ID: <20100621231416.904c50c7.akpm@linux-foundation.org> (raw)
In-Reply-To: <20100622054409.GP7869@dastard>

On Tue, 22 Jun 2010 15:44:09 +1000 Dave Chinner <david@fromorbit.com> wrote:

> > > And so on. This isn't necessarily bad - we'll throttle for longer
> > > than we strictly need to - but the cumulative counter resolution
> > > error gets worse as the number of CPUs doing IO completion grows.
> > > Worst case ends up at for (num cpus * 31) + 1 pages of writeback for
> > > just the first waiter. For an arbitrary FIFO queue of depth d, the
> > > worst case is more like d * (num cpus * 31 + 1).
> >   Hmm, I don't see how the error would depend on the FIFO depth.
> 
> It's the cumulative error that depends on the FIFO depth, not the
> error seen by a single waiter.

Could use the below to basically eliminate the inaccuracies.

Obviously things might get a bit expensive in certain threshold cases
but with some hysteresis that should be manageable.






From: Tim Chen <tim.c.chen@linux.intel.com>

Add percpu_counter_compare that allows for a quick but accurate comparison
of percpu_counter with a given value.

A rough count is provided by the count field in percpu_counter structure,
without accounting for the other values stored in individual cpu counters.
 The actual count is a sum of count and the cpu counters.  However, count
field is never different from the actual value by a factor of
batch*num_online_cpu.  We do not need to get actual count for comparison
if count is different from the given value by this factor and allows for
quick comparison without summing up all the per cpu counters.

Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
---

 include/linux/percpu_counter.h |   11 +++++++++++
 lib/percpu_counter.c           |   27 +++++++++++++++++++++++++++
 2 files changed, 38 insertions(+)

diff -puN include/linux/percpu_counter.h~tmpfs-add-accurate-compare-function-to-percpu_counter-library include/linux/percpu_counter.h
--- a/include/linux/percpu_counter.h~tmpfs-add-accurate-compare-function-to-percpu_counter-library
+++ a/include/linux/percpu_counter.h
@@ -40,6 +40,7 @@ void percpu_counter_destroy(struct percp
 void percpu_counter_set(struct percpu_counter *fbc, s64 amount);
 void __percpu_counter_add(struct percpu_counter *fbc, s64 amount, s32 batch);
 s64 __percpu_counter_sum(struct percpu_counter *fbc);
+int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs);
 
 static inline void percpu_counter_add(struct percpu_counter *fbc, s64 amount)
 {
@@ -98,6 +99,16 @@ static inline void percpu_counter_set(st
 	fbc->count = amount;
 }
 
+static inline int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs)
+{
+	if (fbc->count > rhs)
+		return 1;
+	else if (fbc->count < rhs)
+		return -1;
+	else
+		return 0;
+}
+
 static inline void
 percpu_counter_add(struct percpu_counter *fbc, s64 amount)
 {
diff -puN lib/percpu_counter.c~tmpfs-add-accurate-compare-function-to-percpu_counter-library lib/percpu_counter.c
--- a/lib/percpu_counter.c~tmpfs-add-accurate-compare-function-to-percpu_counter-library
+++ a/lib/percpu_counter.c
@@ -138,6 +138,33 @@ static int __cpuinit percpu_counter_hotc
 	return NOTIFY_OK;
 }
 
+/*
+ * Compare counter against given value.
+ * Return 1 if greater, 0 if equal and -1 if less
+ */
+int percpu_counter_compare(struct percpu_counter *fbc, s64 rhs)
+{
+	s64	count;
+
+	count = percpu_counter_read(fbc);
+	/* Check to see if rough count will be sufficient for comparison */
+	if (abs(count - rhs) > (percpu_counter_batch*num_online_cpus())) {
+		if (count > rhs)
+			return 1;
+		else
+			return -1;
+	}
+	/* Need to use precise count */
+	count = percpu_counter_sum(fbc);
+	if (count > rhs)
+		return 1;
+	else if (count < rhs)
+		return -1;
+	else
+		return 0;
+}
+EXPORT_SYMBOL(percpu_counter_compare);
+
 static int __init percpu_counter_startup(void)
 {
 	compute_batch_value();
_

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2010-06-22  6:14 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-06-17 18:04 [PATCH RFC] mm: Implement balance_dirty_pages() through waiting for flusher thread Jan Kara
2010-06-18  6:09 ` Dave Chinner
2010-06-18  9:11   ` Peter Zijlstra
2010-06-18 23:29     ` Dave Chinner
2010-06-21 23:36   ` Jan Kara
2010-06-22  5:44     ` Dave Chinner
2010-06-22  6:14       ` Andrew Morton [this message]
2010-06-22  7:45         ` Peter Zijlstra
2010-06-22  8:24           ` Andrew Morton
2010-06-22  8:52             ` Peter Zijlstra
2010-06-22 10:09         ` Dave Chinner
2010-06-22 13:17           ` Jan Kara
2010-06-22 13:52             ` Wu Fengguang
2010-06-22 13:59               ` Peter Zijlstra
2010-06-22 14:00               ` Peter Zijlstra
2010-06-22 14:36                 ` Wu Fengguang
2010-06-22 14:02               ` Jan Kara
2010-06-22 14:24                 ` Wu Fengguang
2010-06-22 22:29                 ` Dave Chinner
2010-06-23 13:15                   ` Jan Kara
2010-06-23 23:06                     ` Dave Chinner
2010-06-22 14:31               ` Christoph Hellwig
2010-06-22 14:38                 ` Jan Kara
2010-06-22 22:45                   ` Dave Chinner
2010-06-23  1:34                     ` Wu Fengguang
2010-06-23  3:06                       ` Dave Chinner
2010-06-23  3:22                         ` Wu Fengguang
2010-06-23  6:03                           ` Dave Chinner
2010-06-23  6:25                             ` Wu Fengguang
2010-06-23 23:42                               ` Dave Chinner
2010-06-22 14:41                 ` Wu Fengguang
2010-06-22 11:19       ` Jan Kara
2010-06-18 10:21 ` Peter Zijlstra
2010-06-21 13:31   ` Jan Kara
2010-06-18 10:21 ` Peter Zijlstra
2010-06-21 14:02   ` Jan Kara
2010-06-21 14:10     ` Jan Kara
2010-06-21 14:12       ` Peter Zijlstra
2010-06-18 10:21 ` Peter Zijlstra
2010-06-21 13:42   ` Jan Kara
2010-06-22  4:07     ` Wu Fengguang
2010-06-22 13:27       ` Jan Kara
2010-06-22 13:33         ` Wu Fengguang

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100621231416.904c50c7.akpm@linux-foundation.org \
    --to=akpm@linux-foundation.org \
    --cc=david@fromorbit.com \
    --cc=hch@infradead.org \
    --cc=jack@suse.cz \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=peterz@infradead.org \
    --cc=wfg@mail.ustc.edu.cn \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).