linux-kernel - Re: [PATCH v3 15/15] drm/tests: hdmi: Add test for unsuccessful forced fallback to YUV420

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] [day] [month] [year] [list]

Message-ID: <a4a2680a-eaf7-4505-afbd-b3a71dd46d5a@collabora.com>
Date: Thu, 10 Apr 2025 15:07:40 +0300
From: Cristian Ciocaltea <cristian.ciocaltea@...labora.com>
To: Maxime Ripard <mripard@...nel.org>
Cc: Maarten Lankhorst <maarten.lankhorst@...ux.intel.com>,
 Thomas Zimmermann <tzimmermann@...e.de>, David Airlie <airlied@...il.com>,
 Simona Vetter <simona@...ll.ch>,
 Dave Stevenson <dave.stevenson@...pberrypi.com>,
 Dmitry Baryshkov <dmitry.baryshkov@....qualcomm.com>,
 Dmitry Baryshkov <dmitry.baryshkov@...aro.org>, kernel@...labora.com,
 dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 15/15] drm/tests: hdmi: Add test for unsuccessful
 forced fallback to YUV420

On 4/10/25 12:21 PM, Maxime Ripard wrote:
> Hi,
> 
> On Wed, Mar 26, 2025 at 12:20:04PM +0200, Cristian Ciocaltea wrote:
>> Provide test to verify a forced fallback to YUV420 output cannot succeed
>> when driver doesn't advertise YUV420 support.
>>
>> Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@...labora.com>
>> ---
>>  drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c | 46 ++++++++++++++++++++++
>>  1 file changed, 46 insertions(+)
>>
>> diff --git a/drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c b/drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c
>> index 99bedb2d6f555b3b140256000dfa7491d2a8f515..c2976b42aa2aacd2a68a871bffe97e795ca713d4 100644
>> --- a/drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c
>> +++ b/drivers/gpu/drm/tests/drm_hdmi_state_helper_test.c
>> @@ -1493,6 +1493,51 @@ static void drm_test_check_max_tmds_rate_format_fallback_yuv420(struct kunit *te
>>  	drm_modeset_acquire_fini(&ctx);
>>  }
>>  
>> +/*
>> + * Test that if a driver supports only RGB, but the chosen mode can be
>> + * supported by the screen only in YUV420 output format, we end up with
>> + * an unsuccessful forced fallback attempt.
> 
> What do you mean by "forced"?

I wanted to describe the context where there's no alternative other than
doing the fallback.  Should we replace with "mandatory" maybe?

>> + */
>> +static void drm_test_check_driver_unsupported_fallback_yuv420(struct kunit *test)
>> +{
>> +	struct drm_atomic_helper_connector_hdmi_priv *priv;
>> +	struct drm_modeset_acquire_ctx ctx;
>> +	struct drm_display_info *info;
>> +	struct drm_display_mode *yuv420_only_mode;
>> +	struct drm_connector *conn;
>> +	struct drm_device *drm;
>> +	struct drm_crtc *crtc;
>> +	int ret;
>> +
>> +	priv = drm_kunit_helper_connector_hdmi_init_with_edid(test,
>> +				BIT(HDMI_COLORSPACE_RGB),
>> +				12,
>> +				test_edid_hdmi_1080p_rgb_yuv_4k_yuv420_dc_max_200mhz);
>> +	KUNIT_ASSERT_NOT_NULL(test, priv);
>> +
>> +	drm = &priv->drm;
>> +	crtc = priv->crtc;
>> +	conn = &priv->connector;
>> +	info = &conn->display_info;
>> +	KUNIT_ASSERT_TRUE(test, info->is_hdmi);
>> +	KUNIT_ASSERT_FALSE(test, conn->ycbcr_420_allowed);
>> +
>> +	yuv420_only_mode = drm_kunit_display_mode_from_cea_vic(test, drm, 95);
>> +	KUNIT_ASSERT_NOT_NULL(test, yuv420_only_mode);
>> +	KUNIT_ASSERT_TRUE(test, drm_mode_is_420_only(info, yuv420_only_mode));
>> +
>> +	drm_modeset_acquire_init(&ctx, 0);
>> +
>> +	ret = drm_kunit_helper_enable_crtc_connector(test, drm,
>> +						     crtc, conn,
>> +						     yuv420_only_mode,
>> +						     &ctx);
>> +	KUNIT_EXPECT_LT(test, ret, 0);
> 
> I think that's where your current approach falls a bit short. You should
> really craft the state yourself and check the returned value of
> drm_atomic_check_only(), not rely on
> drm_kunit_helper_enable_crtc_connector() doing the right thing, when it
> doesn't really tell you :)

Ack.

> 
>> +	drm_modeset_drop_locks(&ctx);
>> +	drm_modeset_acquire_fini(&ctx);
>> +}
>> +
> 
> We still need to do the same with a driver that supports both, but the
> monitor doesn't.

Ack.

Thanks,
Cristian