Skip to contentSkip to conversation

Implement Data Access Monitoring-based Memory Operation Schemes

RFC Superseded SeongJae Park 1 participant 8 messages Aug 4, 2020

Subsystems (5)

v14 of 15 Superseded Latest is v15 →

lore
Share & tools
lore
#0 lore raw

v14 cover Implement Data Access Monitoring-based Memory Operation Schemes

From: SeongJae Park <[email protected]>

Changes from Previous Version
=============================

- Drop loadable module support
- Use dedicated valid action checker function
- Rebase on v5.8 plus v19 DAMON

Introduction
============

DAMON[1] can be used as a primitive for data access awared memory management optimizations. For that, users who want such optimizations should run DAMON, read the monitoring results, analyze it, plan a new memory management scheme, and apply the new scheme by themselves. Such efforts will be inevitable for some complicated optimizations.

However, in many other cases, the users would simply want the system to apply a memory management action to a memory region of a specific size having a specific access frequency for a specific time. For example, "page out a memory region larger than 100 MiB keeping only rare accesses more than 2 minutes", or "Do not use THP for a memory region larger than 2 MiB rarely accessed for more than 1 seconds".

This RFC patchset makes DAMON to handle such data access monitoring-based operation schemes. With this change, users can do the data access aware optimizations by simply specifying their schemes to DAMON.

[1] https://lore.kernel.org/linux-mm/[email protected]/

Evaluations
===========

We evaluated DAMON's overhead, monitoring quality and usefulness using 25 realistic workloads on my QEMU/KVM based virtual machine running a kernel that v12 of this patchset is applied.

An experimental DAMON-based operation scheme for THP, ‘ethp’, removes 31.29% of THP memory overheads while preserving 60.64% of THP speedup. Another experimental DAMON-based ‘proactive reclamation’ implementation, ‘prcl’, reduces 87.95% of residential sets and 29.52% of system memory footprint while incurring only 2.15% runtime overhead in the best case (parsec3/freqmine).

NOTE that the experimentail THP optimization and proactive reclamation are not for production, just only for proof of concepts.

Please refer to the official document[1] or "Documentation/admin-guide/mm: Add a document for DAMON" patch in the latest DAMON patchset for detailed evaluation setup and results.

[1] https://damonitor.github.io/doc/html/latest-damos

More Information
================

We prepared a showcase web site[1] that you can get more information. There are

- the official documentations[2],
- the heatmap format dynamic access pattern of various realistic workloads for heap area[3], mmap()-ed area[4], and stack[5] area,
- the dynamic working set size distribution[6] and chronological working set size changes[7], and
- the latest performance test results[8].

[1] https://damonitor.github.io/_index
[2] https://damonitor.github.io/doc/html/latest-damos
[3] https://damonitor.github.io/test/result/visual/latest/rec.heatmap.0.html
[4] https://damonitor.github.io/test/result/visual/latest/rec.heatmap.1.html
[5] https://damonitor.github.io/test/result/visual/latest/rec.heatmap.2.html
[6] https://damonitor.github.io/test/result/visual/latest/rec.wss_sz.html
[7] https://damonitor.github.io/test/result/visual/latest/rec.wss_time.html
[8] https://damonitor.github.io/test/result/perf/latest/html/index.html

Baseline and Complete Git Tree
==============================

The patches are based on the v5.8 plus v19 DAMON patchset[1] and Minchan's ``do_madvise()`` patch[2], which retrieved from the -next tree. You can also clone the complete git tree:

    $ git clone git://github.com/sjp38/linux -b damos/rfc/v14

The web is also available:
https://github.com/sjp38/linux/releases/tag/damos/rfc/v14

There are a couple of trees for entire DAMON patchset series that future features are included. The first one[3] contains the changes for latest release, while the other one[4] contains the changes for next release.

[1] https://lore.kernel.org/linux-mm/[email protected]/
[2] https://lore.kernel.org/linux-mm/[email protected]/
[3] https://github.com/sjp38/linux/tree/damon/master
[4] https://github.com/sjp38/linux/tree/damon/next

Sequence Of Patches
===================

The 1st patch accounts age of each region. The 2nd patch implements the handling of the schemes in DAMON and exports a kernel space programming interface for it. The 3rd patch implements a debugfs interface for the privileged people and user programs. The 4th patch implements schemes statistics feature for easier tuning of the schemes and runtime access pattern analysis. The 5th patche adds selftests for these changes, and the 6th patch adds human friendly schemes support to the user space tool for DAMON. Finally, the 7th patch documents this new feature in the document.

Patch History
=============

Changes from RFC v13
(https://lore.kernel.org/linux-mm/[email protected]/)
- Drop loadable module support
- Use dedicated valid action checker function
- Rebase on v5.8 plus v19 DAMON

Changes from RFC v12
(https://lore.kernel.org/linux-mm/[email protected]/)
- Wordsmith the document, comment, commit messages
- Support a scheme of max access count 0
- Use 'unsigned long' for (min|max)_sz_region

Changes from RFC v11
(https://lore.kernel.org/linux-mm/[email protected]/)
- Refine the commit messages (David Hildenbrand)
- Clean up debugfs code

Changes from RFC v10
(https://lore.kernel.org/linux-mm/[email protected]/)
- Fix the wrong error handling for schemes debugfs file
- Handle the schemes stats from the user space tool
- Remove the schemes implementation plan from the document

Changes from RFC v9
(https://lore.kernel.org/linux-mm/[email protected]/)
- Rebase on v5.7
- Fix wrong comments and documents for schemes apply conditions

Changes from RFC v8
(https://lore.kernel.org/linux-mm/[email protected]/)
- Rewrite the document (Stefan Nuernberger)
- Make 'damon_for_each_*' argument order consistent (Leonard Foerster)
- Implement statistics for schemes
- Avoid races between debugfs readers and writers
- Reset age for only significant access frequency changes
- Add kernel-doc comments in damon.h

Please refer to RFC v8 for previous history

SeongJae Park (7):
mm/damon: Account age of target regions
mm/damon: Implement data access monitoring-based operation schemes mm/damon/schemes: Implement a debugfs interface mm/damon/schemes: Implement statistics feature mm/damon/selftests: Add 'schemes' debugfs tests damon/tools: Support more human friendly 'schemes' control Docs/admin-guide/mm/damon: Document DAMON-based Operation Schemes

 Documentation/admin-guide/mm/damon/guide.rst  |  41 +-
 Documentation/admin-guide/mm/damon/start.rst  |  11 +
 Documentation/admin-guide/mm/damon/usage.rst  | 109 +++++-
 Documentation/vm/damon/index.rst              |   1 -
 include/linux/damon.h                         |  64 ++++
 mm/damon.c                                    | 360 +++++++++++++++++-
 tools/damon/_convert_damos.py                 | 141 +++++++
 tools/damon/_damon.py                         |  28 +-
 tools/damon/damo                              |   7 +
 tools/damon/schemes.py                        | 110 ++++++
 .../testing/selftests/damon/debugfs_attrs.sh  |  29 ++
 11 files changed, 884 insertions(+), 17 deletions(-)
 create mode 100755 tools/damon/_convert_damos.py
 create mode 100644 tools/damon/schemes.py
Signature · 1 line
2.17.1
#1 lore raw

RFC v14 1/7 mm/damon: Account age of target regions

From: SeongJae Park <[email protected]>

DAMON can be used as a primitive for data access pattern aware memory management optimizations. However, users who want such optimizations should run DAMON, read the monitoring results, analyze it, plan a new memory management scheme, and apply the new scheme by themselves. It would not be too hard, but still require some level of effort. For complicated optimizations, this effort is inevitable.

That said, in many cases, users would simply want to apply an actions to a memory region of a specific size having a specific access frequency for a specific time. For example, "page out a memory region larger than 100 MiB but having a low access frequency more than 10 minutes", or "Use THP for a memory region larger than 2 MiB having a high access frequency for more than 2 seconds".

For such optimizations, users will need to first account the age of each region themselves. To reduce such efforts, this commit implements a simple age account of each region in DAMON. For each aggregation step, DAMON compares the access frequency with that from last aggregation and reset the age of the region if the change is significant. Else, the age is incremented. Also, in case of the merge of regions, the region size-weighted average of the ages is set as the age of merged new region.

---
 include/linux/damon.h | 10 ++++++++++
 mm/damon.c            | 20 +++++++++++++++++---
 2 files changed, 27 insertions(+), 3 deletions(-)
include/linux/damon.h +10 -0
index ad771c39d1a9..ca5bd8de4aa4 100644
@@ -31,12 +31,22 @@ struct damon_addr_range {
  * @sampling_addr:	Address of the sample for the next access check.
  * @nr_accesses:	Access frequency of this region.
  * @list:		List head for siblings.
+ * @age:		Age of this region.
+ * @last_nr_accesses:	Internal value for age calculation.
+ *
+ * @age is initially zero, increased for each aggregation interval, and reset
+ * to zero again if the access frequency is significantly changed.  If two
+ * regions are merged into a new region, both @nr_accesses and @age of the new
+ * region are set as region size-weighted average of those of the two regions.
  */
 struct damon_region {
 	struct damon_addr_range ar;
 	unsigned long sampling_addr;
 	unsigned int nr_accesses;
 	struct list_head list;
+
+	unsigned int age;
+	unsigned int last_nr_accesses;
 };
 
 /**
mm/damon.c +17 -3
index 114ef319eeab..69090aa22207 100644
@@ -107,6 +107,9 @@ static struct damon_region *damon_new_region(unsigned long start,
 	region->nr_accesses = 0;
 	INIT_LIST_HEAD(&region->list);
 
+	region->age = 0;
+	region->last_nr_accesses = 0;
+
 	return region;
 }
 
@@ -881,6 +884,7 @@ static void kdamond_reset_aggregated(struct damon_ctx *c)
 			damon_write_rbuf(c, &r->nr_accesses,
 					sizeof(r->nr_accesses));
 			trace_damon_aggregated(t, r, nr);
+			r->last_nr_accesses = r->nr_accesses;
 			r->nr_accesses = 0;
 		}
 	}
@@ -894,9 +898,11 @@ static void kdamond_reset_aggregated(struct damon_ctx *c)
 static void damon_merge_two_regions(struct damon_region *l,
 				struct damon_region *r)
 {
-	l->nr_accesses = (l->nr_accesses * sz_damon_region(l) +
-			r->nr_accesses * sz_damon_region(r)) /
-			(sz_damon_region(l) + sz_damon_region(r));
+	unsigned long sz_l = sz_damon_region(l), sz_r = sz_damon_region(r);
+
+	l->nr_accesses = (l->nr_accesses * sz_l + r->nr_accesses * sz_r) /
+			(sz_l + sz_r);
+	l->age = (l->age * sz_l + r->age * sz_r) / (sz_l + sz_r);
 	l->ar.end = r->ar.end;
 	damon_destroy_region(r);
 }
@@ -916,6 +922,11 @@ static void damon_merge_regions_of(struct damon_target *t, unsigned int thres,
 	struct damon_region *r, *prev = NULL, *next;
 
 	damon_for_each_region_safe(r, next, t) {
+		if (diff_of(r->nr_accesses, r->last_nr_accesses) > thres)
+			r->age = 0;
+		else
+			r->age++;
+
 		if (prev && prev->ar.end == r->ar.start &&
 		    diff_of(prev->nr_accesses, r->nr_accesses) <= thres &&
 		    sz_damon_region(prev) + sz_damon_region(r) <= sz_limit)
@@ -959,6 +970,9 @@ static void damon_split_region_at(struct damon_ctx *ctx,
 	new = damon_new_region(r->ar.start + sz_r, r->ar.end);
 	r->ar.end = new->ar.start;
 
+	new->age = r->age;
+	new->last_nr_accesses = r->last_nr_accesses;
+
 	damon_insert_region(new, r, damon_next_region(r));
 }
 
Signature · 1 line
2.17.1
#2 lore raw

RFC v14 2/7 mm/damon: Implement data access monitoring-based operation schemes

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

In many cases, users might use DAMON for simple data access aware memory management optimizations such as applying an operation scheme to a memory region of a specific size having a specific access frequency for a specific time. For example, "page out a memory region larger than 100 MiB but having a low access frequency more than 10 minutes", or "Use THP for a memory region larger than 2 MiB having a high access frequency for more than 2 seconds".

Most simple form of the solution would be doing offline data access pattern profiling using DAMON and modifying the application source code or system configuration based on the profiling results. Or, developing a daemon constructed with two modules (one for access monitoring and the other for applying memory management actions via mlock(), madvise(), sysctl, etc) is imaginable.

To minimize users from spending their time for implementation of such simple data access monitoring-based operation schemes, this commit makes DAMON to handle such schemes directly. With this commit, users can simply specify their desired schemes to DAMON. Then, DAMON will automatically apply the schemes to the the user-specified target processes.

Each of the schemes is composed with conditions for filtering of the target memory regions and desired memory management action for the target. Specifically, the format is::

<min/max size> <min/max access frequency> <min/max age> <action>

The filtering conditions are size of memory region, number of accesses to the region monitored by DAMON, and the age of the region. The age of region is incremented periodically but reset when its addresses or access frequency has significantly changed or the action of a scheme was applied. For the action, current implementation supports only a few of madvise() hints, ``MADV_WILLNEED``, ``MADV_COLD``, ``MADV_PAGEOUT``, ``MADV_HUGEPAGE``, and ``MADV_NOHUGEPAGE``.

---
 include/linux/damon.h |  48 ++++++++++++++
 mm/damon.c            | 146 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 194 insertions(+)
include/linux/damon.h +48 -0
index ca5bd8de4aa4..df10bd00443e 100644
@@ -67,6 +67,50 @@ struct damon_target {
 	struct list_head list;
 };
 
+/**
+ * enum damos_action - Represents an action of a Data Access Monitoring-based
+ * Operation Scheme.
+ *
+ * @DAMOS_WILLNEED:	Call ``madvise()`` for the region with MADV_WILLNEED.
+ * @DAMOS_COLD:		Call ``madvise()`` for the region with MADV_COLD.
+ * @DAMOS_PAGEOUT:	Call ``madvise()`` for the region with MADV_PAGEOUT.
+ * @DAMOS_HUGEPAGE:	Call ``madvise()`` for the region with MADV_HUGEPAGE.
+ * @DAMOS_NOHUGEPAGE:	Call ``madvise()`` for the region with MADV_NOHUGEPAGE.
+ */
+enum damos_action {
+	DAMOS_WILLNEED,
+	DAMOS_COLD,
+	DAMOS_PAGEOUT,
+	DAMOS_HUGEPAGE,
+	DAMOS_NOHUGEPAGE,
+};
+
+/**
+ * struct damos - Represents a Data Access Monitoring-based Operation Scheme.
+ * @min_sz_region:	Minimum size of target regions.
+ * @max_sz_region:	Maximum size of target regions.
+ * @min_nr_accesses:	Minimum ``->nr_accesses`` of target regions.
+ * @max_nr_accesses:	Maximum ``->nr_accesses`` of target regions.
+ * @min_age_region:	Minimum age of target regions.
+ * @max_age_region:	Maximum age of target regions.
+ * @action:		&damo_action to be applied to the target regions.
+ * @list:		List head for siblings.
+ *
+ * For each aggregation interval, DAMON applies @action to monitoring target
+ * regions fit in the condition and updates the statistics.  Note that both
+ * the minimums and the maximums are inclusive.
+ */
+struct damos {
+	unsigned long min_sz_region;
+	unsigned long max_sz_region;
+	unsigned int min_nr_accesses;
+	unsigned int max_nr_accesses;
+	unsigned int min_age_region;
+	unsigned int max_age_region;
+	enum damos_action action;
+	struct list_head list;
+};
+
 /**
  * struct damon_ctx - Represents a context for each monitoring.  This is the
  * main interface that allows users to set the attributes and get the results
@@ -113,6 +157,7 @@ struct damon_target {
  * @kdamond_lock.  Accesses to other fields must be protected by themselves.
  *
  * @targets_list:	Head of monitoring targets (&damon_target) list.
+ * @schemes_list:	Head of schemes (&damos) list.
  *
  * @init_target_regions:	Constructs initial monitoring target regions.
  * @update_target_regions:	Updates monitoring target regions.
@@ -167,6 +212,7 @@ struct damon_ctx {
 	struct mutex kdamond_lock;
 
 	struct list_head targets_list;	/* 'damon_target' objects */
+	struct list_head schemes_list;	/* 'damos' objects */
 
 	/* callbacks */
 	void (*init_target_regions)(struct damon_ctx *context);
@@ -190,6 +236,8 @@ int damon_set_targets(struct damon_ctx *ctx,
 int damon_set_attrs(struct damon_ctx *ctx, unsigned long sample_int,
 		unsigned long aggr_int, unsigned long regions_update_int,
 		unsigned long min_nr_reg, unsigned long max_nr_reg);
+int damon_set_schemes(struct damon_ctx *ctx,
+			struct damos **schemes, ssize_t nr_schemes);
 int damon_set_recording(struct damon_ctx *ctx,
 				unsigned int rbuf_len, char *rfile_path);
 int damon_start(struct damon_ctx *ctx);
mm/damon.c +146 -0

188 lines not on this page · raw message

Signature · 1 line
2.17.1
#3 lore raw

RFC v14 3/7 mm/damon/schemes: Implement a debugfs interface

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

This commit implements a debugfs interface for the data access monitoring oriented memory management schemes. It is supposed to be used by administrators and/or privileged user space programs. Users can read and update the rules using ``<debugfs>/damon/schemes`` file. The format is::

<min/max size> <min/max access frequency> <min/max age> <action>

---
 mm/damon.c | 187 ++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 185 insertions(+), 2 deletions(-)
mm/damon.c +185 -2
index e402717a2c0e..d6b4181c5d70 100644
@@ -194,6 +194,29 @@ static void damon_destroy_target(struct damon_target *t)
 	damon_free_target(t);
 }
 
+static struct damos *damon_new_scheme(
+		unsigned long min_sz_region, unsigned long max_sz_region,
+		unsigned int min_nr_accesses, unsigned int max_nr_accesses,
+		unsigned int min_age_region, unsigned int max_age_region,
+		enum damos_action action)
+{
+	struct damos *scheme;
+
+	scheme = kmalloc(sizeof(*scheme), GFP_KERNEL);
+	if (!scheme)
+		return NULL;
+	scheme->min_sz_region = min_sz_region;
+	scheme->max_sz_region = max_sz_region;
+	scheme->min_nr_accesses = min_nr_accesses;
+	scheme->max_nr_accesses = max_nr_accesses;
+	scheme->min_age_region = min_age_region;
+	scheme->max_age_region = max_age_region;
+	scheme->action = action;
+	INIT_LIST_HEAD(&scheme->list);
+
+	return scheme;
+}
+
 static void damon_add_scheme(struct damon_ctx *ctx, struct damos *s)
 {
 	list_add_tail(&s->list, &ctx->schemes_list);
@@ -1540,6 +1563,160 @@ static ssize_t debugfs_monitor_on_write(struct file *file,
 	return ret;
 }
 
+static ssize_t sprint_schemes(struct damon_ctx *c, char *buf, ssize_t len)
+{
+	struct damos *s;
+	int written = 0;
+	int rc;
+
+	damon_for_each_scheme(s, c) {
+		rc = snprintf(&buf[written], len - written,
+				"%lu %lu %u %u %u %u %d\n",
+				s->min_sz_region, s->max_sz_region,
+				s->min_nr_accesses, s->max_nr_accesses,
+				s->min_age_region, s->max_age_region,
+				s->action);
+		if (!rc)
+			return -ENOMEM;
+
+		written += rc;
+	}
+	return written;
+}
+
+static ssize_t debugfs_schemes_read(struct file *file, char __user *buf,
+		size_t count, loff_t *ppos)
+{
+	struct damon_ctx *ctx = &damon_user_ctx;
+	char *kbuf;
+	ssize_t len;
+
+	kbuf = kmalloc(count, GFP_KERNEL);
+	if (!kbuf)
+		return -ENOMEM;
+
+	mutex_lock(&ctx->kdamond_lock);
+	len = sprint_schemes(ctx, kbuf, count);
+	mutex_unlock(&ctx->kdamond_lock);
+	if (len < 0)
+		goto out;
+	len = simple_read_from_buffer(buf, count, ppos, kbuf, len);
+
+out:
+	kfree(kbuf);
+	return len;
+}
+
+static void free_schemes_arr(struct damos **schemes, ssize_t nr_schemes)
+{
+	ssize_t i;
+
+	for (i = 0; i < nr_schemes; i++)
+		kfree(schemes[i]);
+	kfree(schemes);
+}
+
+static bool damos_action_valid(int action)
+{
+	switch (action) {
+	case DAMOS_WILLNEED:
+	case DAMOS_COLD:
+	case DAMOS_PAGEOUT:
+	case DAMOS_HUGEPAGE:
+	case DAMOS_NOHUGEPAGE:
+	case DAMOS_STAT:
+		return true;
+	default:
+		return false;
+	}
+}
+
+/*
+ * Converts a string into an array of struct damos pointers
+ *
+ * Returns an array of struct damos pointers that converted if the conversion
+ * success, or NULL otherwise.
+ */
+static struct damos **str_to_schemes(const char *str, ssize_t len,
+				ssize_t *nr_schemes)
+{
+	struct damos *scheme, **schemes;
+	const int max_nr_schemes = 256;
+	int pos = 0, parsed, ret;
+	unsigned long min_sz, max_sz;
+	unsigned int min_nr_a, max_nr_a, min_age, max_age;
+	unsigned int action;
+
+	schemes = kmalloc_array(max_nr_schemes, sizeof(scheme),
+			GFP_KERNEL);
+	if (!schemes)
+		return NULL;
+
+	*nr_schemes = 0;
+	while (pos < len && *nr_schemes < max_nr_schemes) {
+		ret = sscanf(&str[pos], "%lu %lu %u %u %u %u %u%n",
+				&min_sz, &max_sz, &min_nr_a, &max_nr_a,
+				&min_age, &max_age, &action, &parsed);
+		if (ret != 7)
+			break;
+		if (!damos_action_valid(action)) {
+			pr_err("wrong action %d\n", action);
+			goto fail;
+		}
+
+		pos += parsed;
+		scheme = damon_new_scheme(min_sz, max_sz, min_nr_a, max_nr_a,
+				min_age, max_age, action);
+		if (!scheme)
+			goto fail;
+
+		schemes[*nr_schemes] = scheme;
+		*nr_schemes += 1;
+	}
+	return schemes;
+fail:
+	free_schemes_arr(schemes, *nr_schemes);
+	return NULL;
+}
+
+static ssize_t debugfs_schemes_write(struct file *file, const char __user *buf,
+		size_t count, loff_t *ppos)
+{
+	struct damon_ctx *ctx = &damon_user_ctx;
+	char *kbuf;
+	struct damos **schemes;
+	ssize_t nr_schemes = 0, ret = count;
+	int err;
+
+	kbuf = user_input_str(buf, count, ppos);
+	if (IS_ERR(kbuf))
+		return PTR_ERR(kbuf);
+
+	schemes = str_to_schemes(kbuf, ret, &nr_schemes);
+	if (!schemes) {
+		ret = -EINVAL;
+		goto out;
+	}
+
+	mutex_lock(&ctx->kdamond_lock);
+	if (ctx->kdamond) {
+		ret = -EBUSY;
+		goto unlock_out;
+	}
+
+	err = damon_set_schemes(ctx, schemes, nr_schemes);
+	if (err)
+		ret = err;
+	else
+		nr_schemes = 0;
+unlock_out:
+	mutex_unlock(&ctx->kdamond_lock);
+	free_schemes_arr(schemes, nr_schemes);
+out:
+	kfree(kbuf);
+	return ret;
+}
+
 #define targetid_is_pid(ctx)	\
 	(ctx->target_valid == kdamond_vm_target_valid)
 
@@ -1808,6 +1985,12 @@ static const struct file_operations target_ids_fops = {
 	.write = debugfs_target_ids_write,
 };
 
+static const struct file_operations schemes_fops = {
+	.owner = THIS_MODULE,
+	.read = debugfs_schemes_read,
+	.write = debugfs_schemes_write,
+};
+
 static const struct file_operations record_fops = {
 	.owner = THIS_MODULE,
 	.read = debugfs_record_read,
@@ -1824,10 +2007,10 @@ static struct dentry *debugfs_root;
 
 static int __init damon_debugfs_init(void)
 {
-	const char * const file_names[] = {"attrs", "record",
+	const char * const file_names[] = {"attrs", "record", "schemes",
 		"target_ids", "monitor_on"};
 	const struct file_operations *fops[] = {&attrs_fops, &record_fops,
-		&target_ids_fops, &monitor_on_fops};
+		&schemes_fops, &target_ids_fops, &monitor_on_fops};
 	int i;
 
 	debugfs_root = debugfs_create_dir("damon", NULL);
Signature · 1 line
2.17.1
#4 lore raw

RFC v14 4/7 mm/damon/schemes: Implement statistics feature

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

To tune the DAMON-based operation schemes, knowing how many and how large regions are affected by each of the schemes will be helful. Those stats could be used for not only the tuning, but also monitoring of the working set size and the number of regions, if the scheme does not change the program behavior too much.

For the reason, this commit implements the statistics for the schemes. The total number and size of the regions that each scheme is applied are exported to users via '->stat_count' and '->stat_sz' of 'struct damos'. Admins can also check the number by reading 'schemes' debugfs file. The last two integers now represents the stats. To allow collecting the stats without changing the program behavior, this commit also adds new scheme action, 'DAMOS_STAT'. Note that 'DAMOS_STAT' is not only making no memory operation actions, but also does not reset the age of regions.

---
 include/linux/damon.h |  6 ++++++
 mm/damon.c            | 13 ++++++++++---
 2 files changed, 16 insertions(+), 3 deletions(-)
include/linux/damon.h +6 -0
index df10bd00443e..f02798ac9ec5 100644
@@ -76,6 +76,7 @@ struct damon_target {
  * @DAMOS_PAGEOUT:	Call ``madvise()`` for the region with MADV_PAGEOUT.
  * @DAMOS_HUGEPAGE:	Call ``madvise()`` for the region with MADV_HUGEPAGE.
  * @DAMOS_NOHUGEPAGE:	Call ``madvise()`` for the region with MADV_NOHUGEPAGE.
+ * @DAMOS_STAT:		Do nothing but count the stat.
  */
 enum damos_action {
 	DAMOS_WILLNEED,
@@ -83,6 +84,7 @@ enum damos_action {
 	DAMOS_PAGEOUT,
 	DAMOS_HUGEPAGE,
 	DAMOS_NOHUGEPAGE,
+	DAMOS_STAT,		/* Do nothing but only record the stat */
 };
 
 /**
@@ -94,6 +96,8 @@ enum damos_action {
  * @min_age_region:	Minimum age of target regions.
  * @max_age_region:	Maximum age of target regions.
  * @action:		&damo_action to be applied to the target regions.
+ * @stat_count:		Total number of regions that this scheme is applied.
+ * @stat_sz:		Total size of regions that this scheme is applied.
  * @list:		List head for siblings.
  *
  * For each aggregation interval, DAMON applies @action to monitoring target
@@ -108,6 +112,8 @@ struct damos {
 	unsigned int min_age_region;
 	unsigned int max_age_region;
 	enum damos_action action;
+	unsigned long stat_count;
+	unsigned long stat_sz;
 	struct list_head list;
 };
 
mm/damon.c +10 -3
index d6b4181c5d70..d25aeccf7939 100644
@@ -212,6 +212,8 @@ static struct damos *damon_new_scheme(
 	scheme->min_age_region = min_age_region;
 	scheme->max_age_region = max_age_region;
 	scheme->action = action;
+	scheme->stat_count = 0;
+	scheme->stat_sz = 0;
 	INIT_LIST_HEAD(&scheme->list);
 
 	return scheme;
@@ -993,6 +995,8 @@ static int damos_do_action(struct damon_target *target, struct damon_region *r,
 	case DAMOS_NOHUGEPAGE:
 		madv_action = MADV_NOHUGEPAGE;
 		break;
+	case DAMOS_STAT:
+		return 0;
 	default:
 		pr_warn("Wrong action %d\n", action);
 		return -EINVAL;
@@ -1017,8 +1021,11 @@ static void damon_do_apply_schemes(struct damon_ctx *c,
 			continue;
 		if (r->age < s->min_age_region || s->max_age_region < r->age)
 			continue;
+		s->stat_count++;
+		s->stat_sz += sz;
 		damos_do_action(t, r, s->action);
-		r->age = 0;
+		if (s->action != DAMOS_STAT)
+			r->age = 0;
 	}
 }
 
@@ -1571,11 +1578,11 @@ static ssize_t sprint_schemes(struct damon_ctx *c, char *buf, ssize_t len)
 
 	damon_for_each_scheme(s, c) {
 		rc = snprintf(&buf[written], len - written,
-				"%lu %lu %u %u %u %u %d\n",
+				"%lu %lu %u %u %u %u %d %lu %lu\n",
 				s->min_sz_region, s->max_sz_region,
 				s->min_nr_accesses, s->max_nr_accesses,
 				s->min_age_region, s->max_age_region,
-				s->action);
+				s->action, s->stat_count, s->stat_sz);
 		if (!rc)
 			return -ENOMEM;
 
Signature · 1 line
2.17.1
#5 lore raw

RFC v14 5/7 mm/damon/selftests: Add 'schemes' debugfs tests

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

This commit adds simple selftets for 'schemes' debugfs file of DAMON.

---
 .../testing/selftests/damon/debugfs_attrs.sh  | 29 +++++++++++++++++++
 1 file changed, 29 insertions(+)
tools/testing/selftests/damon/debugfs_attrs.sh +29 -0
index c75557e8ba58..61fd3e5598e9 100755
@@ -97,6 +97,35 @@ fi
 
 echo $ORIG_CONTENT > $file
 
+# Test schemes file
+file="$DBGFS/schemes"
+
+ORIG_CONTENT=$(cat $file)
+echo "1 2 3 4 5 6 3" > $file
+if [ $? -ne 0 ]
+then
+	echo "$file write fail"
+	echo $ORIG_CONTENT > $file
+	exit 1
+fi
+
+echo "1 2
+3 4 5 6 3" > $file
+if [ $? -eq 0 ]
+then
+	echo "$file multi line write success (expected fail)"
+	echo $ORIG_CONTENT > $file
+	exit 1
+fi
+
+echo > $file
+if [ $? -ne 0 ]
+then
+	echo "$file empty string writing fail"
+	echo $ORIG_CONTENT > $file
+	exit 1
+fi
+
 # Test target_ids file
 file="$DBGFS/target_ids"
 
Signature · 1 line
2.17.1
#6 lore raw

RFC v14 6/7 damon/tools: Support more human friendly 'schemes' control

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

This commit implements 'schemes' subcommand of the damon userspace tool. It can be used to describe and apply the data access monitoring-based operation schemes in more human friendly fashion.

---
 tools/damon/_convert_damos.py | 141 ++++++++++++++++++++++++++++++++++
 tools/damon/_damon.py         |  28 +++++--
 tools/damon/damo              |   7 ++
 tools/damon/schemes.py        | 110 ++++++++++++++++++++++++++
 4 files changed, 280 insertions(+), 6 deletions(-)
 create mode 100755 tools/damon/_convert_damos.py
 create mode 100644 tools/damon/schemes.py
tools/damon/_convert_damos.py new file +141 -0
new file mode 100755
index 000000000000..0fd84b3701c9
@@ -0,0 +1,141 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: GPL-2.0
+
+"""
+Change human readable data access monitoring-based operation schemes to the low
+level input for the '<debugfs>/damon/schemes' file.  Below is an example of the
+schemes written in the human readable format:
+
+    # format is:
+    # <min/max size> <min/max frequency (0-100)> <min/max age> <action>
+    #
+    # B/K/M/G/T for Bytes/KiB/MiB/GiB/TiB
+    # us/ms/s/m/h/d for micro-seconds/milli-seconds/seconds/minutes/hours/days
+    # 'min/max' for possible min/max value.
+
+    # if a region keeps a high access frequency for >=100ms, put the region on
+    # the head of the LRU list (call madvise() with MADV_WILLNEED).
+    min    max      80      max     100ms   max willneed
+
+    # if a region keeps a low access frequency for >=200ms and <=one hour, put
+    # the region on the tail of the LRU list (call madvise() with MADV_COLD).
+    min     max     10      20      200ms   1h  cold
+
+    # if a region keeps a very low access frequency for >=60 seconds, swap out
+    # the region immediately (call madvise() with MADV_PAGEOUT).
+    min     max     0       10      60s     max pageout
+
+    # if a region of a size >=2MiB keeps a very high access frequency for
+    # >=100ms, let the region to use huge pages (call madvise() with
+    # MADV_HUGEPAGE).
+    2M      max     90      100     100ms   max hugepage
+
+    # If a regions of a size >=2MiB keeps small access frequency for >=100ms,
+    # avoid the region using huge pages (call madvise() with MADV_NOHUGEPAGE).
+    2M      max     0       25      100ms   max nohugepage
+"""
+
+import argparse
+import platform
+
+uint_max = 2**32 - 1
+ulong_max = 2**64 - 1
+if platform.architecture()[0] != '64bit':
+    ulong_max = 2**32 - 1
+
+unit_to_bytes = {'B': 1, 'K': 1024, 'M': 1024 * 1024, 'G': 1024 * 1024 * 1024,
+        'T': 1024 * 1024 * 1024 * 1024}
+
+def text_to_bytes(txt):
+    if txt == 'min':
+        return 0
+    if txt == 'max':
+        return ulong_max
+
+    unit = txt[-1]
+    number = float(txt[:-1])
+    return int(number * unit_to_bytes[unit])
+
+unit_to_usecs = {'us': 1, 'ms': 1000, 's': 1000 * 1000, 'm': 60 * 1000 * 1000,
+        'h': 60 * 60 * 1000 * 1000, 'd': 24 * 60 * 60 * 1000 * 1000}
+
+def text_to_aggr_intervals(txt, aggr_interval):
+    if txt == 'min':
+        return 0
+    if txt == 'max':
+        return uint_max
+
+    unit = txt[-2:]
+    if unit in ['us', 'ms']:
+        number = float(txt[:-2])
+    else:
+        unit = txt[-1]
+        number = float(txt[:-1])
+    return int(number * unit_to_usecs[unit]) / aggr_interval
+
+damos_action_to_int = {'DAMOS_WILLNEED': 0, 'DAMOS_COLD': 1,
+        'DAMOS_PAGEOUT': 2, 'DAMOS_HUGEPAGE': 3, 'DAMOS_NOHUGEPAGE': 4,
+        'DAMOS_STAT': 5}
+
+def text_to_damos_action(txt):
+    return damos_action_to_int['DAMOS_' + txt.upper()]
+
+def text_to_nr_accesses(txt, max_nr_accesses):
+    if txt == 'min':
+        return 0
+    if txt == 'max':
+        return max_nr_accesses
+
+    return int(float(txt) * max_nr_accesses / 100)
+
+def debugfs_scheme(line, sample_interval, aggr_interval):
+    fields = line.split()
+    if len(fields) != 7:
+        print('wrong input line: %s' % line)
+        exit(1)
+
+    limit_nr_accesses = aggr_interval / sample_interval
+    try:
+        min_sz = text_to_bytes(fields[0])
+        max_sz = text_to_bytes(fields[1])
+        min_nr_accesses = text_to_nr_accesses(fields[2], limit_nr_accesses)
+        max_nr_accesses = text_to_nr_accesses(fields[3], limit_nr_accesses)
+        min_age = text_to_aggr_intervals(fields[4], aggr_interval)
+        max_age = text_to_aggr_intervals(fields[5], aggr_interval)
+        action = text_to_damos_action(fields[6])
+    except:
+        print('wrong input field')
+        raise
+    return '%d\t%d\t%d\t%d\t%d\t%d\t%d' % (min_sz, max_sz, min_nr_accesses,
+            max_nr_accesses, min_age, max_age, action)
+
+def convert(schemes_file, sample_interval, aggr_interval):
+    lines = []
+    with open(schemes_file, 'r') as f:
+        for line in f:
+            if line.startswith('#'):
+                continue
+            line = line.strip()
+            if line == '':
+                continue
+            lines.append(debugfs_scheme(line, sample_interval, aggr_interval))
+    return '\n'.join(lines)
+
+def main():
+    parser = argparse.ArgumentParser()
+    parser.add_argument('input', metavar='<file>',
+            help='input file describing the schemes')
+    parser.add_argument('-s', '--sample', metavar='<interval>', type=int,
+            default=5000, help='sampling interval (us)')
+    parser.add_argument('-a', '--aggr', metavar='<interval>', type=int,
+            default=100000, help='aggregation interval (us)')
+    args = parser.parse_args()
+
+    schemes_file = args.input
+    sample_interval = args.sample
+    aggr_interval = args.aggr
+
+    print(convert(schemes_file, sample_interval, aggr_interval))
+
+if __name__ == '__main__':
+    main()
tools/damon/_damon.py +22 -6
index 1f6a292e8c25..a4f6c03c23e4 100644
@@ -10,6 +10,7 @@ import subprocess
 
 debugfs_attrs = None
 debugfs_record = None
+debugfs_schemes = None
 debugfs_target_ids = None
 debugfs_monitor_on = None
 
@@ -33,8 +34,9 @@ class Attrs:
     max_nr_regions = None
     rbuf_len = None
     rfile_path = None
+    schemes = None
 
-    def __init__(self, s, a, r, n, x, l, f):
+    def __init__(self, s, a, r, n, x, l, f, c):
         self.sample_interval = s
         self.aggr_interval = a
         self.regions_update_interval = r
@@ -42,12 +44,13 @@ class Attrs:
         self.max_nr_regions = x
         self.rbuf_len = l
         self.rfile_path = f
+        self.schemes = c
 
     def __str__(self):
-        return "%s %s %s %s %s %s %s" % (self.sample_interval,
+        return "%s %s %s %s %s %s %s\n%s" % (self.sample_interval,
                 self.aggr_interval, self.regions_update_interval,
                 self.min_nr_regions, self.max_nr_regions, self.rbuf_len,
-                self.rfile_path)
+                self.rfile_path, self.schemes)
 
     def attr_str(self):
         return "%s %s %s %s %s " % (self.sample_interval, self.aggr_interval,
@@ -66,6 +69,9 @@ class Attrs:
             debugfs_record), shell=True, executable='/bin/bash')
         if ret:
             return ret
+        return subprocess.call('echo %s > %s' % (
+            self.schemes.replace('\n', ' '), debugfs_schemes), shell=True,
+            executable='/bin/bash')
 
 def current_attrs():
     with open(debugfs_attrs, 'r') as f:
@@ -77,17 +83,26 @@ def current_attrs():
     attrs.append(int(rattrs[0]))
     attrs.append(rattrs[1])
 
+    with open(debugfs_schemes, 'r') as f:
+        schemes = f.read()
+
+    # The last two fields in each line are statistics.  Remove those.
+    schemes = [' '.join(x.split()[:-2]) for x in schemes.strip().split('\n')]
+    attrs.append('\n'.join(schemes))
+
     return Attrs(*attrs)
 
 def chk_update_debugfs(debugfs):
     global debugfs_attrs
     global debugfs_record
+    global debugfs_schemes
     global debugfs_target_ids
     global debugfs_monitor_on
 
     debugfs_damon = os.path.join(debugfs, 'damon')
     debugfs_attrs = os.path.join(debugfs_damon, 'attrs')
     debugfs_record = os.path.join(debugfs_damon, 'record')
+    debugfs_schemes = os.path.join(debugfs_damon, 'schemes')
     debugfs_target_ids = os.path.join(debugfs_damon, 'target_ids')
     debugfs_monitor_on = os.path.join(debugfs_damon, 'monitor_on')
 
@@ -95,8 +110,8 @@ def chk_update_debugfs(debugfs):
         print("damon debugfs dir (%s) not found", debugfs_damon)
         exit(1)
 
-    for f in [debugfs_attrs, debugfs_record, debugfs_target_ids,
-            debugfs_monitor_on]:
+    for f in [debugfs_attrs, debugfs_record, debugfs_schemes,
+            debugfs_target_ids, debugfs_monitor_on]:
         if not os.path.isfile(f):
             print("damon debugfs file (%s) not found" % f)
             exit(1)
@@ -112,8 +127,9 @@ def cmd_args_to_attrs(args):
     if not os.path.isabs(args.out):
         args.out = os.path.join(os.getcwd(), args.out)
     rfile_path = args.out
+    schemes = args.schemes
     return Attrs(sample_interval, aggr_interval, regions_update_interval,
-            min_nr_regions, max_nr_regions, rbuf_len, rfile_path)
+            min_nr_regions, max_nr_regions, rbuf_len, rfile_path, schemes)
 
 def set_attrs_argparser(parser):
     parser.add_argument('-d', '--debugfs', metavar='<debugfs>', type=str,
tools/damon/damo +7 -0
index 58e1099ae5fc..ce7180069bef 100755
@@ -5,6 +5,7 @@ import argparse
 
 import record
 import report
+import schemes
 
 class SubCmdHelpFormatter(argparse.RawDescriptionHelpFormatter):
     def _format_action(self, action):
@@ -25,6 +26,10 @@ parser_record = subparser.add_parser('record',
         help='record data accesses of the given target processes')
 record.set_argparser(parser_record)
 
+parser_schemes = subparser.add_parser('schemes',
+        help='apply operation schemes to the given target process')
+schemes.set_argparser(parser_schemes)
+
 parser_report = subparser.add_parser('report',
         help='report the recorded data accesses in the specified form')
 report.set_argparser(parser_report)
@@ -33,5 +38,7 @@ args = parser.parse_args()
 
 if args.command == 'record':
     record.main(args)
+elif args.command == 'schemes':
+    schemes.main(args)
 elif args.command == 'report':
     report.main(args)
tools/damon/schemes.py new file +110 -0

110 lines not on this page · raw message

Signature · 1 line
2.17.1
#7 lore raw

RFC v14 7/7 Docs/admin-guide/mm/damon: Document DAMON-based Operation Schemes

In reply to SeongJae Park

From: SeongJae Park <[email protected]>

This commit add description of DAMON-based operation schemes in the DAMON documents.

---
 Documentation/admin-guide/mm/damon/guide.rst |  41 ++++++-
 Documentation/admin-guide/mm/damon/start.rst |  11 ++
 Documentation/admin-guide/mm/damon/usage.rst | 109 ++++++++++++++++++-
 Documentation/vm/damon/index.rst             |   1 -
 4 files changed, 156 insertions(+), 6 deletions(-)
Documentation/admin-guide/mm/damon/guide.rst +38 -3
index c51fb843efaa..1f9aa2ebbdb6 100644
@@ -53,6 +53,11 @@ heats``.  If it shows a simple pattern consists of a small number of memory
 regions having high contrast of access temperature, you could consider manual
 `Program Modification`_.
 
+If the access pattern is very frequently changing so that you cannot figure out
+what is the performance important region using your human eye, `Automated
+DAMON-based Memory Operations`_ might help the case owing to its machine-level
+microscope view.
+
 If you still want to absorb more benefits, you should develop `Personalized
 DAMON Application`_ for your special case.
 
@@ -120,6 +125,36 @@ shows the visualized access patterns of streamcluster workload in PARSEC3
 benchmark suite.  We can easily identify the 100 MiB sized hot object.
 
 
+Automated DAMON-based Memory Operations
+---------------------------------------
+
+Though `Manual Program Optimization` works well in many cases and DAMON can
+help it, modifying the source code is not a good option in many cases.  First
+of all, the source code could be too old or unavailable.  And, many workloads
+will have complex data access patterns that even hard to distinguish hot memory
+objects and cold memory objects with the human eye.  Finding the mapping from
+the visualized access pattern to the source code and injecting the hinting
+system calls inside the code will also be quite challenging.
+
+By using DAMON-based operation schemes (DAMOS) via ``damo schemes``, you will
+be able to easily optimize your workload in such a case.  Our example schemes
+called 'efficient THP' and 'proactive reclamation' achieved significant speedup
+and memory space saves against 25 realistic workloads [2]_.
+
+That said, note that you need careful tune of the schemes (e.g., target region
+size and age) and monitoring attributes for the successful use of this
+approach.  Because the optimal values of the parameters will be dependent on
+each system and workload, misconfiguring the parameters could result in worse
+memory management.
+
+For the tuning, you could measure the performance metrics such as IPC, TLB
+misses, and swap in/out events and adjusts the parameters based on their
+changes.  The total number and the total size of the regions that each scheme
+is applied, which are provided via the debugfs interface and the programming
+interface can also be useful.  Writing a program automating this optimal
+parameter could be an option.
+
+
 Personalized DAMON Application
 ------------------------------
 
@@ -146,9 +181,9 @@ Referencing previously done successful practices could help you getting the
 sense for this kind of optimizations.  There is an academic paper [1]_
 reporting the visualized access pattern and manual `Program
 Modification`_ results for a number of realistic workloads.  You can also get
-the visualized access patterns [3]_ [4]_ [5]_ and automated DAMON-based memory
-operations results for other realistic workloads that collected with latest
-version of DAMON [2]_ .
+the visualized access patterns [3]_ [4]_ [5]_ and
+`Automated DAMON-based Memory Operations`_ results for other realistic
+workloads that collected with latest version of DAMON [2]_ .
 
 .. [1] https://dl.acm.org/doi/10.1145/3366626.3368125
 .. [2] https://damonitor.github.io/test/result/perf/latest/html/
Documentation/admin-guide/mm/damon/start.rst +11 -0
index deed2ea2321e..35cf4e4ca6aa 100644
@@ -90,6 +90,17 @@ image files. ::
 You can show the images in a web page [1]_ .  Those made with other realistic
 workloads are also available [2]_ [3]_ [4]_.
 
+
+Data Access Pattern Aware Memory Management
+===========================================
+
+Below three commands make every memory region of size >=4K that doesn't
+accessed for >=60 seconds in your workload to be swapped out. ::
+
+    $ echo "#min-size max-size min-acc max-acc min-age max-age action" > scheme
+    $ echo "4K        max      0       0       60s     max     pageout" >> scheme
+    $ damo schemes -c my_thp_scheme <pid of your workload>
+
 .. [1] https://damonitor.github.io/doc/html/v17/admin-guide/mm/damon/start.html#visualizing-recorded-patterns
 .. [2] https://damonitor.github.io/test/result/visual/latest/rec.heatmap.1.png.html
 .. [3] https://damonitor.github.io/test/result/visual/latest/rec.wss_sz.png.html
Documentation/admin-guide/mm/damon/usage.rst +107 -2

124 lines not on this page · raw message

Documentation/vm/damon/index.rst +0 -1
index 17dca3c12aad..69aec1287aaf 100644
@@ -28,4 +28,3 @@ workloads and systems.
    design
    eval
    api
-   plans
Signature · 1 line
2.17.1