linux-kernel - [PATCH 11/11] drm/vc4: tests: pv-muxing: Document test scenario

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <20230710-kms-kunit-actions-rework-v1-11-722c58d72c72@kernel.org>
Date:   Mon, 10 Jul 2023 09:47:42 +0200
From:   Maxime Ripard <mripard@...nel.org>
To:     David Airlie <airlied@...il.com>, Daniel Vetter <daniel@...ll.ch>,
        Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
        Thomas Zimmermann <tzimmermann@...e.de>,
        Emma Anholt <emma@...olt.net>
Cc:     dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org,
        Maxime Ripard <mripard@...nel.org>
Subject: [PATCH 11/11] drm/vc4: tests: pv-muxing: Document test scenario

We've had a couple of tests that weren't really obvious, nor did they
document what they were supposed to test. Document that to make it
hopefully more obvious.

Signed-off-by: Maxime Ripard <mripard@...nel.org>
---
 drivers/gpu/drm/vc4/tests/vc4_test_pv_muxing.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/drivers/gpu/drm/vc4/tests/vc4_test_pv_muxing.c b/drivers/gpu/drm/vc4/tests/vc4_test_pv_muxing.c
index 5f9f5626329d..61622e951031 100644
--- a/drivers/gpu/drm/vc4/tests/vc4_test_pv_muxing.c
+++ b/drivers/gpu/drm/vc4/tests/vc4_test_pv_muxing.c
@@ -845,6 +845,13 @@ static void drm_test_vc5_pv_muxing_bugs_subsequent_crtc_enable(struct kunit *tes
 	KUNIT_EXPECT_NE(test, hdmi0_channel, hdmi1_channel);
 }
 
+/*
+ * This test makes sure that we never change the FIFO of an active HVS
+ * channel if we disable a FIFO with a lower index.
+ *
+ * Doing so would result in a FIFO stall and would disrupt an output
+ * supposed to be unaffected by the commit.
+ */
 static void drm_test_vc5_pv_muxing_bugs_stable_fifo(struct kunit *test)
 {
 	struct drm_modeset_acquire_ctx *ctx;
@@ -924,6 +931,21 @@ static void drm_test_vc5_pv_muxing_bugs_stable_fifo(struct kunit *test)
 	}
 }
 
+/*
+ * Test that if we affect a single output, only the CRTC state of that
+ * output will be pulled in the global atomic state.
+ *
+ * This is relevant for two things:
+ *
+ *   - If we don't have that state at all, we are unlikely to affect the
+ *     FIFO muxing. This is somewhat redundant with
+ *     drm_test_vc5_pv_muxing_bugs_stable_fifo()
+ *
+ *   - KMS waits for page flips to occur on all the CRTC found in the
+ *     CRTC state. Since the CRTC is unaffected, we would over-wait, but
+ *     most importantly run into corner cases like waiting on an
+ *     inactive CRTC that never completes.
+ */
 static void
 drm_test_vc5_pv_muxing_bugs_subsequent_crtc_enable_too_many_crtc_state(struct kunit *test)
 {

-- 
2.41.0