From: Peter Zijlstra <a.p.zijlstra@chello.nl>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Cc: miklos@szeredi.hu, akpm@linux-foundation.org, neilb@suse.de,
dgc@sgi.com, tomoki.sekiyama.qu@hitachi.com,
a.p.zijlstra@chello.nl, nikita@clusterfs.com,
trond.myklebust@fys.uio.no, yingchao.zhou@gmail.com,
richard@rsk.demon.co.uk, torvalds@linux-foundation.org
Subject: [PATCH 22/23] mm: dirty balancing for tasks
Date: Fri, 03 Aug 2007 14:37:35 +0200 [thread overview]
Message-ID: <20070803125237.728389000@chello.nl> (raw)
In-Reply-To: 20070803123712.987126000@chello.nl
[-- Attachment #1: dirty_pages2.patch --]
[-- Type: text/plain, Size: 5322 bytes --]
Based on ideas of Andrew:
http://marc.info/?l=linux-kernel&m=102912915020543&w=2
Scale the bdi dirty limit inversly with the tasks dirty rate.
This makes heavy writers have a lower dirty limit than the occasional writer.
Andrea proposed something similar:
http://lwn.net/Articles/152277/
The main disadvantage to his patch is that he uses an unrelated quantity to
measure time, which leaves him with a workload dependant tunable. Other than
that the two approached appear quite similar.
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
---
include/linux/sched.h | 2 +
kernel/exit.c | 1
kernel/fork.c | 8 +++++++
mm/page-writeback.c | 56 +++++++++++++++++++++++++++++++++++++++++++++++++-
4 files changed, 66 insertions(+), 1 deletion(-)
Index: linux-2.6/include/linux/sched.h
===================================================================
--- linux-2.6.orig/include/linux/sched.h
+++ linux-2.6/include/linux/sched.h
@@ -86,6 +86,7 @@ struct sched_param {
#include <linux/timer.h>
#include <linux/hrtimer.h>
#include <linux/task_io_accounting.h>
+#include <linux/proportions.h>
#include <asm/processor.h>
@@ -1188,6 +1189,7 @@ struct task_struct {
#ifdef CONFIG_FAULT_INJECTION
int make_it_fail;
#endif
+ struct prop_local_single dirties;
};
/*
Index: linux-2.6/kernel/exit.c
===================================================================
--- linux-2.6.orig/kernel/exit.c
+++ linux-2.6/kernel/exit.c
@@ -161,6 +161,7 @@ repeat:
ptrace_unlink(p);
BUG_ON(!list_empty(&p->ptrace_list) || !list_empty(&p->ptrace_children));
__exit_signal(p);
+ prop_local_destroy(&p->dirties);
/*
* If we are the last non-leader member of the thread
Index: linux-2.6/kernel/fork.c
===================================================================
--- linux-2.6.orig/kernel/fork.c
+++ linux-2.6/kernel/fork.c
@@ -163,6 +163,7 @@ static struct task_struct *dup_task_stru
{
struct task_struct *tsk;
struct thread_info *ti;
+ int err;
prepare_to_copy(orig);
@@ -176,6 +177,13 @@ static struct task_struct *dup_task_stru
return NULL;
}
+ err = prop_local_init(&tsk->dirties);
+ if (err) {
+ free_thread_info(ti);
+ free_task_struct(tsk);
+ return NULL;
+ }
+
*tsk = *orig;
tsk->stack = ti;
setup_thread_stack(tsk, orig);
Index: linux-2.6/mm/page-writeback.c
===================================================================
--- linux-2.6.orig/mm/page-writeback.c
+++ linux-2.6/mm/page-writeback.c
@@ -118,6 +118,7 @@ static void background_writeout(unsigned
*
*/
static struct prop_descriptor vm_completions;
+static struct prop_descriptor vm_dirties;
static unsigned long determine_dirtyable_memory(void);
@@ -146,6 +147,7 @@ int dirty_ratio_handler(ctl_table *table
if (ret == 0 && write && vm_dirty_ratio != old_ratio) {
int shift = calc_period_shift();
prop_change_shift(&vm_completions, shift);
+ prop_change_shift(&vm_dirties, shift);
}
return ret;
}
@@ -161,6 +163,16 @@ static void __bdi_writeout_inc(struct ba
prop_put_global(&vm_completions, pg);
}
+static void task_dirty_inc(struct task_struct *tsk)
+{
+ unsigned long flags;
+ struct prop_global *pg = prop_get_global(&vm_dirties);
+ local_irq_save(flags);
+ __prop_inc(pg, &tsk->dirties);
+ local_irq_restore(flags);
+ prop_put_global(&vm_dirties, pg);
+}
+
/*
* Obtain an accurate fraction of the BDI's portion.
*/
@@ -201,6 +213,38 @@ clip_bdi_dirty_limit(struct backing_dev_
*pbdi_dirty = min(*pbdi_dirty, avail_dirty);
}
+void task_dirties_fraction(struct task_struct *tsk,
+ long *numerator, long *denominator)
+{
+ struct prop_global *pg = prop_get_global(&vm_dirties);
+ prop_fraction(pg, &tsk->dirties, numerator, denominator);
+ prop_put_global(&vm_dirties, pg);
+}
+
+/*
+ * scale the dirty limit
+ *
+ * task specific dirty limit:
+ *
+ * dirty -= (dirty/2) * p_{t}
+ */
+void task_dirty_limit(struct task_struct *tsk, long *pdirty)
+{
+ long numerator, denominator;
+ long dirty = *pdirty;
+ long long inv = dirty >> 1;
+
+ task_dirties_fraction(tsk, &numerator, &denominator);
+ inv *= numerator;
+ do_div(inv, denominator);
+
+ dirty -= inv;
+ if (dirty < *pdirty/2)
+ dirty = *pdirty/2;
+
+ *pdirty = dirty;
+}
+
/*
* Work out the current dirty-memory clamping and background writeout
* thresholds.
@@ -307,6 +351,7 @@ get_dirty_limits(long *pbackground, long
*pbdi_dirty = bdi_dirty;
clip_bdi_dirty_limit(bdi, dirty, pbdi_dirty);
+ task_dirty_limit(current, pbdi_dirty);
}
}
@@ -728,6 +773,7 @@ void __init page_writeback_init(void)
shift = calc_period_shift();
prop_descriptor_init(&vm_completions, shift);
+ prop_descriptor_init(&vm_dirties, shift);
}
/**
@@ -1006,7 +1052,7 @@ EXPORT_SYMBOL(redirty_page_for_writepage
* If the mapping doesn't provide a set_page_dirty a_op, then
* just fall through and assume that it wants buffer_heads.
*/
-int fastcall set_page_dirty(struct page *page)
+static int __set_page_dirty(struct page *page)
{
struct address_space *mapping = page_mapping(page);
@@ -1024,6 +1070,14 @@ int fastcall set_page_dirty(struct page
}
return 0;
}
+
+int fastcall set_page_dirty(struct page *page)
+{
+ int ret = __set_page_dirty(page);
+ if (ret)
+ task_dirty_inc(current);
+ return ret;
+}
EXPORT_SYMBOL(set_page_dirty);
/*
--
next prev parent reply other threads:[~2007-08-03 13:04 UTC|newest]
Thread overview: 190+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-03 12:37 [PATCH 00/23] per device dirty throttling -v8 Peter Zijlstra
2007-08-03 12:37 ` [PATCH 01/23] nfs: remove congestion_end() Peter Zijlstra
2007-08-03 12:37 ` [PATCH 02/23] lib: percpu_counter_add Peter Zijlstra
2007-08-03 12:37 ` [PATCH 03/23] lib: percpu_counter variable batch Peter Zijlstra
2007-08-03 12:37 ` [PATCH 04/23] lib: make percpu_counter_add take s64 Peter Zijlstra
2007-08-03 12:37 ` [PATCH 05/23] lib: percpu_counter_set Peter Zijlstra
2007-08-03 12:37 ` [PATCH 06/23] lib: percpu_counter_sum_positive Peter Zijlstra
2007-08-03 12:37 ` [PATCH 07/23] lib: percpu_count_sum() Peter Zijlstra
2007-08-03 12:37 ` [PATCH 08/23] lib: percpu_counter_init error handling Peter Zijlstra
2007-08-03 12:37 ` [PATCH 09/23] lib: percpu_counter_init_irq Peter Zijlstra
2007-08-03 12:37 ` [PATCH 10/23] mm: bdi init hooks Peter Zijlstra
2007-08-03 12:37 ` [PATCH 11/23] containers: " Peter Zijlstra
2007-08-03 12:37 ` [PATCH 12/23] mtd: " Peter Zijlstra
2007-08-03 12:37 ` [PATCH 13/23] mtd: clean up the backing_dev_info usage Peter Zijlstra
2007-08-03 12:37 ` [PATCH 14/23] mtd: give mtdconcat devices their own backing_dev_info Peter Zijlstra
2007-08-03 12:37 ` [PATCH 15/23] mm: scalable bdi statistics counters Peter Zijlstra
2007-08-03 12:37 ` [PATCH 16/23] mm: count reclaimable pages per BDI Peter Zijlstra
2007-08-03 12:37 ` [PATCH 17/23] mm: count writeback " Peter Zijlstra
2007-08-09 19:15 ` Christoph Lameter
2007-08-09 19:23 ` Peter Zijlstra
2007-08-09 19:27 ` Christoph Lameter
2007-08-13 8:36 ` Peter Zijlstra
2007-08-03 12:37 ` [PATCH 18/23] mm: expose BDI statistics in sysfs Peter Zijlstra
2007-08-03 12:37 ` [PATCH 19/23] lib: floating proportions Peter Zijlstra
2007-08-03 12:37 ` [PATCH 20/23] lib: floating proportions _single Peter Zijlstra
2007-08-03 12:37 ` [PATCH 21/23] mm: per device dirty threshold Peter Zijlstra
2007-08-03 12:37 ` Peter Zijlstra [this message]
2007-08-03 12:37 ` [PATCH 23/23] debug: sysfs files for the current ratio/size/total Peter Zijlstra
2007-08-03 22:21 ` [PATCH 00/23] per device dirty throttling -v8 Linus Torvalds
2007-08-04 6:32 ` Ingo Molnar
2007-08-04 7:07 ` Ingo Molnar
2007-08-04 7:44 ` david
2007-08-04 16:01 ` Ray Lee
2007-08-04 17:15 ` david
2007-08-09 5:11 ` david
2007-08-04 10:33 ` Ingo Molnar
2007-08-04 16:17 ` Linus Torvalds
2007-08-04 16:37 ` Ingo Molnar
2007-08-04 16:51 ` Andrew Morton
2007-08-04 16:56 ` Ingo Molnar
2007-08-04 20:23 ` Alan Cox
2007-08-04 17:02 ` Diego Calleja
2007-08-04 17:17 ` Ingo Molnar
2007-08-04 17:38 ` Diego Calleja
2007-08-04 17:51 ` Diego Calleja
2007-08-08 10:43 ` Karel Zak
2007-08-04 17:39 ` Linus Torvalds
2007-08-04 18:08 ` Jeff Garzik
2007-08-04 19:12 ` Jörn Engel
2007-08-04 19:21 ` Ingo Molnar
2007-08-04 19:26 ` Jörn Engel
2007-08-04 19:42 ` Jörn Engel
2007-08-05 20:36 ` Christoph Hellwig
2007-08-06 18:03 ` Chuck Ebbert
2007-08-06 18:53 ` Jeff Garzik
2007-08-06 19:37 ` Alan Cox
2007-08-06 19:46 ` Chuck Ebbert
2007-08-07 7:05 ` Ingo Molnar
2007-08-08 21:10 ` Martin J. Bligh
2007-08-08 21:21 ` Andrew Morton
2007-08-09 0:54 ` Martin Bligh
2007-08-11 23:14 ` Valerie Henson
2007-08-10 0:21 ` Bill Davidsen
2007-08-14 9:57 ` Helge Hafting
2007-08-04 19:47 ` Linus Torvalds
2007-08-04 19:49 ` Linus Torvalds
2007-08-04 20:00 ` Ingo Molnar
2007-08-04 20:11 ` Ingo Molnar
2007-08-04 20:13 ` Arjan van de Ven
2007-08-05 8:18 ` [patch] add noatime/atime boot options, CONFIG_DEFAULT_NOATIME Ingo Molnar
2007-08-04 20:13 ` [PATCH 00/23] per device dirty throttling -v8 Arjan van de Ven
2007-08-04 21:48 ` Theodore Tso
2007-08-05 18:01 ` Arjan van de Ven
2007-08-05 20:34 ` Christoph Hellwig
[not found] ` <fa.7rstQpXif2z9y2n2HD+qxLFnueg@ifi.uio.no>
[not found] ` <fa.6VOZrceT65Vh8CIRIta0zSg2V38@ifi.uio.no>
[not found] ` <fa.xcZCTa5cHDOhrcyXZ2gZbzbu7g0@ifi.uio.no>
[not found] ` <fa.JjZwG90x+07YaOx8h5VLN+9AL/8@ifi.uio.no>
[not found] ` <fa.V9U4mAEXVjNqblhzu7GRmxif7Uw@ifi.uio.no>
[not found] ` <fa.uq0BQtrgp66a08hpsF+vrqXUNC4@ifi.uio.no>
2007-08-15 18:16 ` david.balazic
2007-08-04 20:11 ` [PATCH 00/23] " Alan Cox
2007-08-04 20:28 ` Jeff Garzik
2007-08-04 21:47 ` Alan Cox
2007-08-04 23:51 ` Claudio Martins
2007-08-05 0:49 ` Alan Cox
2007-08-05 7:28 ` Ingo Molnar
2007-08-05 10:29 ` Jakob Oestergaard
2007-08-05 12:46 ` Alan Cox
2007-08-05 12:58 ` Ingo Molnar
2007-08-05 13:29 ` Willy Tarreau
2007-08-06 6:57 ` Ingo Molnar
2007-08-06 13:12 ` Willy Tarreau
2007-08-05 14:46 ` Theodore Tso
2007-08-05 17:55 ` Ingo Molnar
2007-08-05 17:59 ` Jeff Garzik
2007-08-05 18:09 ` Ingo Molnar
2007-08-05 18:08 ` Arjan van de Ven
2007-08-07 21:20 ` Bill Davidsen
2007-08-05 7:18 ` Ingo Molnar
2007-08-07 18:55 ` Bill Davidsen
2007-08-07 19:35 ` Alan Cox
2007-08-08 17:44 ` Bill Davidsen
2007-08-04 20:28 ` Ingo Molnar
2007-08-04 20:34 ` Arjan van de Ven
2007-08-04 21:03 ` Ingo Molnar
2007-08-04 21:51 ` Alan Cox
2007-08-05 7:21 ` Ingo Molnar
2007-08-05 7:29 ` Andrew Morton
2007-08-05 7:39 ` Ingo Molnar
2007-08-05 8:53 ` Willy Tarreau
2007-08-05 14:17 ` Jörn Engel
2007-08-05 18:02 ` Arjan van de Ven
2007-08-05 18:37 ` Jörn Engel
2007-08-05 20:21 ` Jörn Engel
2007-08-05 20:33 ` Andrew Morton
2007-08-05 12:47 ` Alan Cox
2007-08-05 12:56 ` Ingo Molnar
2007-08-05 18:44 ` Dave Jones
2007-08-05 18:58 ` adi
2007-08-06 6:39 ` Ingo Molnar
2007-08-06 15:59 ` Dave Jones
2007-08-06 16:16 ` Ingo Molnar
2007-08-05 7:37 ` Ingo Molnar
2007-08-05 9:04 ` Jeff Garzik
2007-08-05 12:43 ` Alan Cox
2007-08-05 12:54 ` Ingo Molnar
2007-08-05 13:37 ` Alan Cox
2007-08-05 18:08 ` Ingo Molnar
2007-08-05 19:11 ` Alan Cox
2007-08-08 18:22 ` Bill Davidsen
2007-08-08 19:39 ` Jeff Garzik
2007-08-08 20:31 ` Bill Davidsen
2007-08-08 23:18 ` Alan Cox
2007-08-07 19:09 ` Bill Davidsen
2007-08-04 21:48 ` Alan Cox
2007-08-05 7:13 ` Ingo Molnar
2007-08-05 13:22 ` Diego Calleja
2007-08-05 19:03 ` david
2007-08-06 6:52 ` Ingo Molnar
2007-08-10 4:04 ` Bill Davidsen
2007-08-11 5:19 ` Valdis.Kletnieks
2007-08-06 6:58 ` Ingo Molnar
2007-08-09 0:57 ` Greg Trounson
2007-08-09 1:26 ` david
2007-08-09 2:33 ` Andi Kleen
2007-08-04 22:39 ` Ilpo Järvinen
2007-08-05 10:20 ` Jakob Oestergaard
2007-08-05 10:42 ` Jeff Garzik
2007-08-05 10:58 ` Jakob Oestergaard
2007-08-05 12:46 ` Ingo Molnar
2007-08-05 13:46 ` Jakob Oestergaard
2007-08-05 16:45 ` Linus Torvalds
2007-08-05 19:09 ` Ingo Molnar
2007-08-05 19:22 ` [patch] implement smarter atime updates support Ingo Molnar
2007-08-05 19:28 ` [patch] implement smarter atime updates support, v2 Ingo Molnar
2007-08-05 20:42 ` Theodore Tso
2007-08-06 5:36 ` Ingo Molnar
2007-08-05 19:53 ` [patch] implement smarter atime updates support Arjan van de Ven
2007-08-05 20:04 ` Alan Cox
2007-08-05 20:22 ` Arjan van de Ven
2007-08-05 19:29 ` [PATCH 00/23] per device dirty throttling -v8 Alan Cox
2007-08-05 19:32 ` Ingo Molnar
2007-08-05 23:43 ` David Chinner
2007-08-05 0:26 ` Andi Kleen
2007-08-05 15:00 ` Theodore Tso
2007-08-06 13:47 ` Chris Mason
2007-08-17 0:45 ` Dave Jones
2007-08-05 20:41 ` Christoph Hellwig
2007-08-06 10:42 ` Andi Kleen
2007-08-16 10:18 ` Helge Hafting
2007-08-09 6:25 ` Lionel Elie Mamane
2007-08-09 15:02 ` Chuck Ebbert
2007-08-09 16:22 ` Diego Calleja
2007-08-04 16:41 ` Andrew Morton
2007-08-04 17:26 ` Nikita Danilov
2007-08-04 19:16 ` Florian Weimer
2007-08-05 6:00 ` Andrew Morton
2007-08-05 7:57 ` Florian Weimer
2007-08-05 20:43 ` Christoph Hellwig
2007-08-05 22:46 ` Theodore Tso
2007-08-06 0:24 ` David Chinner
2007-08-05 0:28 ` Andi Kleen
2007-08-04 16:15 ` Linus Torvalds
2007-08-05 17:22 ` Brice Figureau
2007-08-05 22:17 ` Andi Kleen
2007-08-06 8:40 ` Brice Figureau
2007-08-14 1:44 ` Stewart Smith
2007-08-14 2:25 ` Andi Kleen
2007-08-14 7:59 ` Brice Figureau
2007-08-06 20:26 ` Miklos Szeredi
2007-08-08 12:25 ` richard kennedy
2007-08-08 13:54 ` Andi Kleen
2007-08-10 4:17 ` Bill Davidsen
2007-08-16 7:45 [PATCH 00/23] per device dirty throttling -v9 Peter Zijlstra
2007-08-16 7:45 ` [PATCH 22/23] mm: dirty balancing for tasks Peter Zijlstra
2007-09-11 19:53 [PATCH 00/23] per device dirty throttling -v10 Peter Zijlstra
2007-09-11 19:54 ` [PATCH 22/23] mm: dirty balancing for tasks Peter Zijlstra
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=20070803125237.728389000@chello.nl \
--to=a.p.zijlstra@chello.nl \
--cc=akpm@linux-foundation.org \
--cc=dgc@sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=miklos@szeredi.hu \
--cc=neilb@suse.de \
--cc=nikita@clusterfs.com \
--cc=richard@rsk.demon.co.uk \
--cc=tomoki.sekiyama.qu@hitachi.com \
--cc=torvalds@linux-foundation.org \
--cc=trond.myklebust@fys.uio.no \
--cc=yingchao.zhou@gmail.com \
/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).