[RFC PATCH 0/3] mm/damon/sysfs-schemes: Do DAMOS tried regions update for one apply interval

[RFC PATCH 0/3] mm/damon/sysfs-schemes: Do DAMOS tried regions update for one apply interval

[SeongJae Park]

DAMOS tried regions update feature of DAMON sysfs interface is doing the
update for one aggregation interval after the request is made.  Since
the per-scheme apply interval is supported, that behavior makes no much
sense.  That is, the tried regions directory will have regions from
multiple DAMON monitoring results snapshots, or no region for apply
intervals that much shorter than, or longer than the aggregation
interval, respectively.  Update the behavior to update the regions for
each scheme for only its apply interval, and update the document.

Since DAMOS apply interval is the aggregation by default, this change
makes no visible behavioral difference to old users who don't explicitly
set the apply intervals.

Patches Sequence
----------------

The first two patches makes schemes of apply intervals that much shorter
or longer than the aggregation interval to keep the maximum and minimum
times for continuing the update.  After the two patches, the update
aligns with the each scheme's apply interval.

Finally, the third patch updates the document to reflect the behavior.

SeongJae Park (3):
  mm/damon/sysfs-schemes: do not update tried regions more than one
    DAMON snapshot
  mm/damon/sysfs: avoid empty scheme tried regions for large apply
    interval
  Docs/admin-guide/mm/damon/usage: update for tried regions update time
    interval

 Documentation/admin-guide/mm/damon/usage.rst |  6 +-
 mm/damon/sysfs-common.h                      |  2 +
 mm/damon/sysfs-schemes.c                     | 93 ++++++++++++++++++++
 mm/damon/sysfs.c                             | 34 ++++++-
 4 files changed, 128 insertions(+), 7 deletions(-)


base-commit: 369359a808616f7f87172568da7933f65834caab
-- 
2.34.1

[RFC PATCH 1/3] mm/damon/sysfs-schemes: do not update tried regions more than one DAMON snapshot

[SeongJae Park]

DAMON_SYSFS exposes every DAMON-found region that eligible for applying
the scheme action for one aggregation interval.  However, each
DAMON-based operation scheme has its own apply interval.  Hence, for a
scheme that having its apply interval much smaller than the aggregation
interval, DAMON_SYSFS will expose the scheme regions that applied to
more than one DAMON monitoring results snapshots.  Since the purpose of
DAMON tried regions is exposing single snapshot, this makes no much
sense.  Track progress of each scheme's tried regions update and avoid
the case.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 mm/damon/sysfs-schemes.c | 77 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 77 insertions(+)

diff --git a/mm/damon/sysfs-schemes.c b/mm/damon/sysfs-schemes.c
index a7d70b95c4dd..b07a5c544b34 100644
--- a/mm/damon/sysfs-schemes.c
+++ b/mm/damon/sysfs-schemes.c
@@ -113,11 +113,47 @@ static const struct kobj_type damon_sysfs_scheme_region_ktype = {
  * scheme regions directory
  */
 
+/*
+ * enum damos_sysfs_regions_upd_status - Represent DAMOS tried regions update
+ *					 status
+ * @DAMOS_TRIED_REGIONS_UPD_IDLE:		Waiting for next request.
+ * @DAMOS_TRIED_REGIONS_UPD_STARTED:		Update started.
+ * @DAMOS_TRIED_REGIONS_UPD_FINISHED:	Update finished.
+ *
+ * Each DAMON-based operation scheme (&struct damos) has its own apply
+ * interval, and we need to expose the scheme tried regions based on only
+ * single snapshot.  For this, we keep the tried regions update status for each
+ * scheme.  The status becomes 'idle' at the beginning.
+ *
+ * Once the tried regions update request is received, the request handling
+ * start function (damon_sysfs_scheme_update_regions_start()) sets the status
+ * of all schemes as 'idle' again, and register ->before_damos_apply() and
+ * ->after_sampling() callbacks.
+ *
+ * Then, the first followup ->before_damos_apply() callback
+ * (damon_sysfs_before_damos_apply()) sets the status 'started'.  The first
+ * ->after_sampling() callback (damon_sysfs_after_sampling()) after the call
+ * is called only after the scheme is completely applied
+ * to the given snapshot.  Hence the callback knows the situation by showing
+ * 'started' status, and sets the status as 'finished'.  Then,
+ * damon_sysfs_before_damos_apply() understands the situation by showing the
+ * 'finished' status and do nothing.
+ *
+ *  Finally, the tried regions request handling finisher function
+ *  (damon_sysfs_schemes_update_regions_stop()) unregisters the callbacks.
+ */
+enum damos_sysfs_regions_upd_status {
+	DAMOS_TRIED_REGIONS_UPD_IDLE,
+	DAMOS_TRIED_REGIONS_UPD_STARTED,
+	DAMOS_TRIED_REGIONS_UPD_FINISHED,
+};
+
 struct damon_sysfs_scheme_regions {
 	struct kobject kobj;
 	struct list_head regions_list;
 	int nr_regions;
 	unsigned long total_bytes;
+	enum damos_sysfs_regions_upd_status upd_status;
 };
 
 static struct damon_sysfs_scheme_regions *
@@ -130,6 +166,7 @@ damon_sysfs_scheme_regions_alloc(void)
 	INIT_LIST_HEAD(&regions->regions_list);
 	regions->nr_regions = 0;
 	regions->total_bytes = 0;
+	regions->upd_status = DAMOS_TRIED_REGIONS_UPD_IDLE;
 	return regions;
 }
 
@@ -1777,6 +1814,10 @@ static int damon_sysfs_before_damos_apply(struct damon_ctx *ctx,
 		return 0;
 
 	sysfs_regions = sysfs_schemes->schemes_arr[schemes_idx]->tried_regions;
+	if (sysfs_regions->upd_status == DAMOS_TRIED_REGIONS_UPD_FINISHED)
+		return 0;
+	if (sysfs_regions->upd_status == DAMOS_TRIED_REGIONS_UPD_IDLE)
+		sysfs_regions->upd_status = DAMOS_TRIED_REGIONS_UPD_STARTED;
 	sysfs_regions->total_bytes += r->ar.end - r->ar.start;
 	if (damos_regions_upd_total_bytes_only)
 		return 0;
@@ -1793,6 +1834,29 @@ static int damon_sysfs_before_damos_apply(struct damon_ctx *ctx,
 	return 0;
 }
 
+/*
+ * DAMON callback that called after each accesses sampling.  While this
+ * callback is registered, damon_sysfs_lock should be held to ensure the
+ * regions directories exist.
+ */
+static int damon_sysfs_after_sampling(struct damon_ctx *ctx)
+{
+	struct damon_sysfs_schemes *sysfs_schemes =
+		damon_sysfs_schemes_for_damos_callback;
+	struct damon_sysfs_scheme_regions *sysfs_regions;
+	int i;
+
+	for (i = 0; i < sysfs_schemes->nr; i++) {
+		sysfs_regions = sysfs_schemes->schemes_arr[i]->tried_regions;
+		if (sysfs_regions->upd_status ==
+				DAMOS_TRIED_REGIONS_UPD_STARTED)
+			sysfs_regions->upd_status =
+				DAMOS_TRIED_REGIONS_UPD_FINISHED;
+	}
+
+	return 0;
+}
+
 /* Called from damon_sysfs_cmd_request_callback under damon_sysfs_lock */
 int damon_sysfs_schemes_clear_regions(
 		struct damon_sysfs_schemes *sysfs_schemes,
@@ -1816,6 +1880,16 @@ int damon_sysfs_schemes_clear_regions(
 	return 0;
 }
 
+static void damos_tried_regions_init_upd_status(
+		struct damon_sysfs_schemes *sysfs_schemes)
+{
+	int i;
+
+	for (i = 0; i < sysfs_schemes->nr; i++)
+		sysfs_schemes->schemes_arr[i]->tried_regions->upd_status =
+			DAMOS_TRIED_REGIONS_UPD_IDLE;
+}
+
 /* Called from damon_sysfs_cmd_request_callback under damon_sysfs_lock */
 int damon_sysfs_schemes_update_regions_start(
 		struct damon_sysfs_schemes *sysfs_schemes,
@@ -1823,8 +1897,10 @@ int damon_sysfs_schemes_update_regions_start(
 {
 	damon_sysfs_schemes_clear_regions(sysfs_schemes, ctx);
 	damon_sysfs_schemes_for_damos_callback = sysfs_schemes;
+	damos_tried_regions_init_upd_status(sysfs_schemes);
 	damos_regions_upd_total_bytes_only = total_bytes_only;
 	ctx->callback.before_damos_apply = damon_sysfs_before_damos_apply;
+	ctx->callback.after_sampling = damon_sysfs_after_sampling;
 	return 0;
 }
 
@@ -1837,6 +1913,7 @@ int damon_sysfs_schemes_update_regions_stop(struct damon_ctx *ctx)
 {
 	damon_sysfs_schemes_for_damos_callback = NULL;
 	ctx->callback.before_damos_apply = NULL;
+	ctx->callback.after_sampling = NULL;
 	damon_sysfs_schemes_region_idx = 0;
 	return 0;
 }
-- 
2.34.1

[RFC PATCH 2/3] mm/damon/sysfs: avoid empty scheme tried regions for large apply interval

[SeongJae Park]

DAMON_SYSFS assumes all schemes will be applied for at least one DAMON
monitoring results snapshot within one aggregation interval, or makes no
sense to wait for it while DAMON is deactivated by the watermarks.  That
for deactivated status still makes sense, but the aggregation interval
based assumption is invalid now because each scheme can has its own
apply interval.  For schemes having larger than the aggregation or
watermarks check interval, DAMOS tried regions update request can be
finished without the update.  Avoid the case by explicitly checking the
status of the schemes tried regions update and watermarks based DAMON
deactivation.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 mm/damon/sysfs-common.h  |  2 ++
 mm/damon/sysfs-schemes.c | 16 ++++++++++++++++
 mm/damon/sysfs.c         | 34 ++++++++++++++++++++++++++++++----
 3 files changed, 48 insertions(+), 4 deletions(-)

diff --git a/mm/damon/sysfs-common.h b/mm/damon/sysfs-common.h
index fd482a0639b4..5ff081226e28 100644
--- a/mm/damon/sysfs-common.h
+++ b/mm/damon/sysfs-common.h
@@ -49,6 +49,8 @@ int damon_sysfs_schemes_update_regions_start(
 		struct damon_sysfs_schemes *sysfs_schemes,
 		struct damon_ctx *ctx, bool total_bytes_only);
 
+bool damos_sysfs_regions_upd_done(void);
+
 int damon_sysfs_schemes_update_regions_stop(struct damon_ctx *ctx);
 
 int damon_sysfs_schemes_clear_regions(
diff --git a/mm/damon/sysfs-schemes.c b/mm/damon/sysfs-schemes.c
index b07a5c544b34..45bd0fd4a8b1 100644
--- a/mm/damon/sysfs-schemes.c
+++ b/mm/damon/sysfs-schemes.c
@@ -1904,6 +1904,22 @@ int damon_sysfs_schemes_update_regions_start(
 	return 0;
 }
 
+bool damos_sysfs_regions_upd_done(void)
+{
+	struct damon_sysfs_schemes *sysfs_schemes =
+		damon_sysfs_schemes_for_damos_callback;
+	struct damon_sysfs_scheme_regions *sysfs_regions;
+	int i;
+
+	for (i = 0; i < sysfs_schemes->nr; i++) {
+		sysfs_regions = sysfs_schemes->schemes_arr[i]->tried_regions;
+		if (sysfs_regions->upd_status !=
+				DAMOS_TRIED_REGIONS_UPD_FINISHED)
+			return false;
+	}
+	return true;
+}
+
 /*
  * Called from damon_sysfs_cmd_request_callback under damon_sysfs_lock.  Caller
  * should unlock damon_sysfs_lock which held before
diff --git a/mm/damon/sysfs.c b/mm/damon/sysfs.c
index e6b8e90bd450..71bdc51bea6e 100644
--- a/mm/damon/sysfs.c
+++ b/mm/damon/sysfs.c
@@ -1336,12 +1336,13 @@ static int damon_sysfs_commit_input(struct damon_sysfs_kdamond *kdamond)
 
 /*
  * damon_sysfs_cmd_request_callback() - DAMON callback for handling requests.
- * @c:	The DAMON context of the callback.
+ * @c:		The DAMON context of the callback.
+ * @active:	Whether @c is not deactivated due to watermarks.
  *
  * This function is periodically called back from the kdamond thread for @c.
  * Then, it checks if there is a waiting DAMON sysfs request and handles it.
  */
-static int damon_sysfs_cmd_request_callback(struct damon_ctx *c)
+static int damon_sysfs_cmd_request_callback(struct damon_ctx *c, bool active)
 {
 	struct damon_sysfs_kdamond *kdamond;
 	bool total_bytes_only = false;
@@ -1373,6 +1374,13 @@ static int damon_sysfs_cmd_request_callback(struct damon_ctx *c)
 				goto keep_lock_out;
 			}
 		} else {
+			/*
+			 * Continue regions updating if DAMON is till
+			 * active and the update for all schemes is not
+			 * finished.
+			 */
+			if (active && !damos_sysfs_regions_upd_done())
+				goto keep_lock_out;
 			err = damon_sysfs_upd_schemes_regions_stop(kdamond);
 			damon_sysfs_schemes_regions_updating = false;
 		}
@@ -1392,6 +1400,24 @@ static int damon_sysfs_cmd_request_callback(struct damon_ctx *c)
 	return err;
 }
 
+static int damon_sysfs_after_wmarks_check(struct damon_ctx *c)
+{
+	/*
+	 * after_wmarks_check() is called back while the context is deactivated
+	 * by watermarks.
+	 */
+	return damon_sysfs_cmd_request_callback(c, false);
+}
+
+static int damon_sysfs_after_aggregation(struct damon_ctx *c)
+{
+	/*
+	 * after_aggregation() is called back only while the context is not
+	 * deactivated by watermarks.
+	 */
+	return damon_sysfs_cmd_request_callback(c, true);
+}
+
 static struct damon_ctx *damon_sysfs_build_ctx(
 		struct damon_sysfs_context *sys_ctx)
 {
@@ -1407,8 +1433,8 @@ static struct damon_ctx *damon_sysfs_build_ctx(
 		return ERR_PTR(err);
 	}
 
-	ctx->callback.after_wmarks_check = damon_sysfs_cmd_request_callback;
-	ctx->callback.after_aggregation = damon_sysfs_cmd_request_callback;
+	ctx->callback.after_wmarks_check = damon_sysfs_after_wmarks_check;
+	ctx->callback.after_aggregation = damon_sysfs_after_aggregation;
 	ctx->callback.before_terminate = damon_sysfs_before_terminate;
 	return ctx;
 }
-- 
2.34.1

[RFC PATCH 3/3] Docs/admin-guide/mm/damon/usage: update for tried regions update time interval

[SeongJae Park]

The documentation says DAMOS tried regions update feature of DAMON sysfs
interface is doing the update for one aggregation interval after the
request is made.  Since the introduction of the per-scheme apply
interval, that behavior makes no much sense.  Hence the implementation
has changed to update the regions for each scheme for only its apply
interval.  Further update the document to reflect the real behavior.

Signed-off-by: SeongJae Park <sj@kernel.org>
---
 Documentation/admin-guide/mm/damon/usage.rst | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/Documentation/admin-guide/mm/damon/usage.rst b/Documentation/admin-guide/mm/damon/usage.rst
index 8507a6e45d86..da94feb97ed1 100644
--- a/Documentation/admin-guide/mm/damon/usage.rst
+++ b/Documentation/admin-guide/mm/damon/usage.rst
@@ -432,9 +432,9 @@ that reading it returns the total size of the scheme tried regions, and creates
 directories named integer starting from ``0`` under this directory.  Each
 directory contains files exposing detailed information about each of the memory
 region that the corresponding scheme's ``action`` has tried to be applied under
-this directory, during next :ref:`aggregation interval
-<sysfs_monitoring_attrs>`.  The information includes address range,
-``nr_accesses``, and ``age`` of the region.
+this directory, during next :ref:`apply interval <damon_design_damos>` of the
+corresponding scheme.  The information includes address range, ``nr_accesses``,
+and ``age`` of the region.
 
 Writing ``update_schemes_tried_bytes`` to the relevant ``kdamonds/<N>/state``
 file will only update the ``total_bytes`` file, and will not create the
-- 
2.34.1