Skip to contentSkip to conversation

mm/damon/tests/core-kunit: add test for damon_call() with repeat control

PATCH Superseded Sailesh Nandanavanam 2 participants 3 messages Nov 17 – Nov 18

Subsystems (2)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 2/2 mm/damon/tests/core-kunit: add test for damon_call() with repeat control

When call_controls_obsolete is unset and damon_call_control->repeat is set, damon_call() should queue the request and return 0 immediately without waiting for kdamond to handle it. Add a kunit test for that.

---
 mm/damon/tests/core-kunit.h | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
mm/damon/tests/core-kunit.h +25 -0
index 13315291f69f..bc93677487b5 100644
@@ -2499,6 +2499,30 @@ static void damon_test_call_controls_obsolete(struct kunit *test)
 	damon_destroy_ctx(ctx);
 }
 
+static void damon_test_call_repeat(struct kunit *test)
+{
+	struct damon_ctx *ctx;
+	struct damon_call_control control = {};
+	int ret;
+
+	ctx = damon_new_ctx();
+
+	if (!ctx)
+		kunit_skip(test, "ctx alloc fail");
+
+	ctx->call_controls_obsolete = false;
+	control.repeat = true;
+
+	ret = damon_call(ctx, &control);
+
+	KUNIT_EXPECT_EQ(test, ret, 0);
+	KUNIT_EXPECT_FALSE(test, list_empty(&ctx->call_controls));
+	KUNIT_EXPECT_PTR_EQ(test, ctx->call_controls.next, &control.list);
+
+	list_del(&control.list);
+	damon_destroy_ctx(ctx);
+}
+
 static struct kunit_case damon_test_cases[] = {
 	KUNIT_CASE(damon_test_target),
 	KUNIT_CASE(damon_test_regions),
@@ -2544,6 +2568,7 @@ static struct kunit_case damon_test_cases[] = {
 	KUNIT_CASE(damos_test_esz_goal_temporal),
 	KUNIT_CASE(damon_test_last_region_quota_reset),
 	KUNIT_CASE(damon_test_call_controls_obsolete),
+	KUNIT_CASE(damon_test_call_repeat),
 	{},
 };
 
Signature · 1 line
2.34.1
maintainer
#1 lore raw
Quoted 3 lines · On Tue, 17 Nov 2026 23:18:34 +0530 Sailesh Nandanavanam…

When call_controls_obsolete is unset and damon_call_control->repeat is set, damon_call() should queue the request and return 0 immediately without waiting for kdamond to handle it. Add a kunit test for that.

Apparently this patch was meant to be sent as a part of a series with another patch [1] but individually sent by a setup failure. As a result, this patch is not reviewed by Sashiko. Could you please repost as the series?

Quoted 23 lines
---
 mm/damon/tests/core-kunit.h | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)
mm/damon/tests/core-kunit.h +9 -0
index 13315291f69f..bc93677487b5 100644
@@ -2499,6 +2499,30 @@ static void damon_test_call_controls_obsolete(struct kunit *test)
 	damon_destroy_ctx(ctx);
 }
 
+static void damon_test_call_repeat(struct kunit *test)
+{
+	struct damon_ctx *ctx;
+	struct damon_call_control control = {};
+	int ret;
+
+	ctx = damon_new_ctx();
+
+	if (!ctx)

mm/damon/tests/core-kunit.h static void damon_test_call_controls_obsolete(struct kunit *test) · line 2520

Quoted 10 lines
mm/damon/tests/core-kunit.h +10 -0
+		kunit_skip(test, "ctx alloc fail");
+
+	ctx->call_controls_obsolete = false;
+	control.repeat = true;
+
+	ret = damon_call(ctx, &control);
+
+	KUNIT_EXPECT_EQ(test, ret, 0);
+	KUNIT_EXPECT_FALSE(test, list_empty(&ctx->call_controls));
+	KUNIT_EXPECT_PTR_EQ(test, ctx->call_controls.next, &control.list);

I think the above last two expects are somewhat too much implementation detail. I don't think we are promising those list status. I'd prefer dropping those.

mm/damon/tests/core-kunit.h static void damon_test_call_controls_obsolete(struct kunit *test) · line 2522

Quoted 2 lines
mm/damon/tests/core-kunit.h +2 -0
+
+	list_del(&control.list);

Can anything leak or corrupted if we don't call list_del() here?

mm/damon/tests/core-kunit.h static struct kunit_case damon_test_cases[] = { · line 2571

Quoted 11 lines
mm/damon/tests/core-kunit.h +4 -0
+	damon_destroy_ctx(ctx);
+}
+
 static struct kunit_case damon_test_cases[] = {
 	KUNIT_CASE(damon_test_target),
 	KUNIT_CASE(damon_test_regions),
@@ -2544,6 +2568,7 @@ static struct kunit_case damon_test_cases[] = {
 	KUNIT_CASE(damos_test_esz_goal_temporal),
 	KUNIT_CASE(damon_test_last_region_quota_reset),
 	KUNIT_CASE(damon_test_call_controls_obsolete),
+	KUNIT_CASE(damon_test_call_repeat),

Could we merge the above two tests into one, say, damon_test_call()?

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

Thanks,
SJ

[...]

#2 lore raw

From: Sailesh Nandanavanam <[email protected]>

Quoted 2 lines · On Wed, 18 Nov 2026 SJ Park <[email protected]> wrote:

I think the above last two expects are somewhat too much implementation detail. I don't think we are promising those list status. I'd prefer dropping those.

Agreed, dropping both list assertions.

Quoted 1 line

Can anything leak or corrupted if we don't call list_del() here?

No. The control is stack-allocated and damon_destroy_ctx() does not touch call_controls, so nothing is leaked or accessed afterwards. Will drop the list_del().

Quoted 1 line

Could we merge the above two tests into one, say, damon_test_call()?

Will merge them into damon_test_call() and send v2 as a single patch.

Quoted 3 lines

Apparently this patch was meant to be sent as a part of a series with another patch [1] but individually sent by a setup failure. As a result, this patch is not reviewed by Sashiko. Could you please repost as the series?

Sorry for the mess. Since the two tests will be merged, v2 will be a single patch, so Sashiko will review it.

Thanks,
Sailesh