Skip to main content

Browse the archive

Revision d7831a0bdf06b9f722b947bb0c205ff7d77cebd8 authored by Richard Kennedy on 30 June 2009, 18:41:35 UTC, committed by Linus Torvalds on 01 July 2009, 01:56:01 UTC 
mm: prevent balance_dirty_pages() from doing too much work
 
balance_dirty_pages can overreact and move all of the dirty pages to
writeback unnecessarily.

balance_dirty_pages makes its decision to throttle based on the number of
dirty plus writeback pages that are over the calculated limit,so it will
continue to move pages even when there are plenty of pages in writeback
and less than the threshold still dirty.

This allows it to overshoot its limits and move all the dirty pages to
writeback while waiting for the drives to catch up and empty the writeback
list.

A simple fio test easily demonstrates this problem.

fio --name=f1 --directory=/disk1 --size=2G -rw=write --name=f2 --directory=/disk2 --size=1G --rw=write --startdelay=10

This is the simplest fix I could find, but I'm not entirely sure that it
alone will be enough for all cases.  But it certainly is an improvement on
my desktop machine writing to 2 disks.

Do we need something more for machines with large arrays where
bdi_threshold * number_of_drives is greater than the dirty_ratio ?

Signed-off-by: Richard Kennedy <richard@rsk.demon.co.uk>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
1 parent df279ca
Raw File
dm-snap-transient.c 
/*
 * Copyright (C) 2001-2002 Sistina Software (UK) Limited.
 * Copyright (C) 2006-2008 Red Hat GmbH
 *
 * This file is released under the GPL.
 */

#include "dm-exception-store.h"

#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/vmalloc.h>
#include <linux/slab.h>
#include <linux/dm-io.h>

#define DM_MSG_PREFIX "transient snapshot"

/*-----------------------------------------------------------------
 * Implementation of the store for non-persistent snapshots.
 *---------------------------------------------------------------*/
struct transient_c {
	sector_t next_free;
};

static void transient_dtr(struct dm_exception_store *store)
{
	kfree(store->context);
}

static int transient_read_metadata(struct dm_exception_store *store,
				   int (*callback)(void *callback_context,
						   chunk_t old, chunk_t new),
				   void *callback_context)
{
	return 0;
}

static int transient_prepare_exception(struct dm_exception_store *store,
				       struct dm_snap_exception *e)
{
	struct transient_c *tc = store->context;
	sector_t size = get_dev_size(store->cow->bdev);

	if (size < (tc->next_free + store->chunk_size))
		return -1;

	e->new_chunk = sector_to_chunk(store, tc->next_free);
	tc->next_free += store->chunk_size;

	return 0;
}

static void transient_commit_exception(struct dm_exception_store *store,
				       struct dm_snap_exception *e,
				       void (*callback) (void *, int success),
				       void *callback_context)
{
	/* Just succeed */
	callback(callback_context, 1);
}

static void transient_fraction_full(struct dm_exception_store *store,
				    sector_t *numerator, sector_t *denominator)
{
	*numerator = ((struct transient_c *) store->context)->next_free;
	*denominator = get_dev_size(store->cow->bdev);
}

static int transient_ctr(struct dm_exception_store *store,
			 unsigned argc, char **argv)
{
	struct transient_c *tc;

	tc = kmalloc(sizeof(struct transient_c), GFP_KERNEL);
	if (!tc)
		return -ENOMEM;

	tc->next_free = 0;
	store->context = tc;

	return 0;
}

static unsigned transient_status(struct dm_exception_store *store,
				 status_type_t status, char *result,
				 unsigned maxlen)
{
	unsigned sz = 0;

	switch (status) {
	case STATUSTYPE_INFO:
		break;
	case STATUSTYPE_TABLE:
		DMEMIT(" %s N %llu", store->cow->name,
		       (unsigned long long)store->chunk_size);
	}

	return sz;
}

static struct dm_exception_store_type _transient_type = {
	.name = "transient",
	.module = THIS_MODULE,
	.ctr = transient_ctr,
	.dtr = transient_dtr,
	.read_metadata = transient_read_metadata,
	.prepare_exception = transient_prepare_exception,
	.commit_exception = transient_commit_exception,
	.fraction_full = transient_fraction_full,
	.status = transient_status,
};

static struct dm_exception_store_type _transient_compat_type = {
	.name = "N",
	.module = THIS_MODULE,
	.ctr = transient_ctr,
	.dtr = transient_dtr,
	.read_metadata = transient_read_metadata,
	.prepare_exception = transient_prepare_exception,
	.commit_exception = transient_commit_exception,
	.fraction_full = transient_fraction_full,
	.status = transient_status,
};

int dm_transient_snapshot_init(void)
{
	int r;

	r = dm_exception_store_type_register(&_transient_type);
	if (r) {
		DMWARN("Unable to register transient exception store type");
		return r;
	}

	r = dm_exception_store_type_register(&_transient_compat_type);
	if (r) {
		DMWARN("Unable to register old-style transient "
		       "exception store type");
		dm_exception_store_type_unregister(&_transient_type);
		return r;
	}

	return r;
}

void dm_transient_snapshot_exit(void)
{
	dm_exception_store_type_unregister(&_transient_type);
	dm_exception_store_type_unregister(&_transient_compat_type);
}
back to top