All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Nautiyal, Ankit K" <ankit.k.nautiyal@intel.com>
To: "Peres, Martin" <martin.peres@intel.com>,
	"igt-dev@lists.freedesktop.org" <igt-dev@lists.freedesktop.org>
Cc: "Latvala, Petri" <petri.latvala@intel.com>
Subject: Re: [igt-dev] [PATCH i-g-t] tests/kms_content_protection: Bifurcate tests into HDCP1.4 & 2.2
Date: Mon, 23 Mar 2020 10:34:22 +0530	[thread overview]
Message-ID: <aae888e6-b9fc-5ffb-57f5-645b7dad8b19@intel.com> (raw)
In-Reply-To: <414609087dad419caa67e95008d3c2ec@intel.com>



On 3/20/2020 9:33 PM, Peres, Martin wrote:
> On 2020-03-20 11:03, Ankit Nautiyal wrote:
>> Currently, the test kms_content_protection does not distinguish
>> between HDCP1.4 and HDCP2.2 tests. Since the driver by default
>> always chooses HDCP2.2 protocol if the panel supports HDCP1.4 &
>> HDCP2.2, HDCP1.4 does not get tested.
> Thanks for working on it!
>
>> This patch makes two versions of the existing subtests for testing
>> HDCP1.4 and HDCP2.2. For forcing the driver to use HDCP1.4, the
>> kernel module 'mei_hdcp' responsible for HDCP2.2 authentication,
>> is unloaded before HDCP1.4 tests. Once the HDCP1.4 tests are
>> completed, the module is re-loaded.
>>
>> This patch also makes the existing subtests as dynamic sub-subtests
>> based on all outputs supporting the protocol under test.
>> This helps in running together all HDCP1.4 tests for all outputs
>> that support HDCP1.4, avoiding the need of loading and unloading
>> of 'mei_hdcp' module for each test.
> That sounds like a complicated process which won't scale to future
> versions of HDCP (no reason to assume they will stop pumping them out).
>
> How about adding a debugfs file that lists the supported hdcp versions
> for every connector, says which one is being used right now, and
> force-disable one or many so as to be able to validate all the cases?
>
> Martin

If I understand correctly, the debugfs will tell two things:

i) The currently selected version for a platform say 'global_hdcp_version'.
The platform will always go with this version of HDCP. So by default it 
is HDCP2.2 for a platform (say KBL), but we can select to use HDCP1.4.

ii) The list of connectors supporting the selected platforms.
So if global_hdcp_version is set to be HDCP2.2, only connectors that 
support HDCP2.2 will be listed.

The test will then become simpler. Reading from the debugfs and start 
with lowest version going to highest supported.
1) Set the global_hdcp_version to hdcp_version[i]
2) Loop for each connector, listed for selected version.
3) set appropriate connector properties.
4) Validate.

The work of making sure, that selected version is chosen, will then lie 
with the driver.
So for KBL/ICL/TGL etc,  default is HDCP2.2, setting to HDCP1.4, it 
might need to unload mei_hdcp.
For next platforms and version, the work might be something different.

I hope I get it right to some extent :) Thanks for the comments.


Regards,
Ankit


>> Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
>> ---
>>   tests/kms_content_protection.c | 185 ++++++++++++++++++++-------------
>>   1 file changed, 113 insertions(+), 72 deletions(-)
>>
>> diff --git a/tests/kms_content_protection.c b/tests/kms_content_protection.c
>> index 3b9cedcb..1472e987 100644
>> --- a/tests/kms_content_protection.c
>> +++ b/tests/kms_content_protection.c
>> @@ -59,6 +59,9 @@ struct data {
>>   #define HDCP_CONTENT_TYPE_0				0
>>   #define HDCP_CONTENT_TYPE_1				1
>>   
>> +#define HDCP_14						0
>> +#define HDCP_22						1
>> +
>>   #define LIC_PERIOD_MSEC				(4 * 1000)
>>   /* Kernel retry count=3, Max time per authentication allowed = 6Sec */
>>   #define KERNEL_AUTH_TIME_ALLOWED_MSEC		(3 *  6 * 1000)
>> @@ -593,44 +596,6 @@ static bool sink_hdcp2_capable(igt_output_t *output)
>>   	return strstr(buf, "HDCP2.2");
>>   }
>>   
>> -static void
>> -test_content_protection(enum igt_commit_style s, int content_type)
>> -{
>> -	igt_display_t *display = &data.display;
>> -	igt_output_t *output;
>> -	int valid_tests = 0;
>> -
>> -	if (data.cp_tests & CP_MEI_RELOAD)
>> -		igt_require_f(igt_kmod_is_loaded("mei_hdcp"),
>> -			      "mei_hdcp module is not loaded\n");
>> -
>> -	for_each_connected_output(display, output) {
>> -		if (!output->props[IGT_CONNECTOR_CONTENT_PROTECTION])
>> -			continue;
>> -
>> -		if (!output->props[IGT_CONNECTOR_HDCP_CONTENT_TYPE] &&
>> -		    content_type)
>> -			continue;
>> -
>> -		igt_info("CP Test execution on %s\n", output->name);
>> -
>> -		if (content_type && !sink_hdcp2_capable(output)) {
>> -			igt_info("\tSkip %s (Sink has no HDCP2.2 support)\n",
>> -				 output->name);
>> -			continue;
>> -		} else if (!sink_hdcp_capable(output)) {
>> -			igt_info("\tSkip %s (Sink has no HDCP support)\n",
>> -				 output->name);
>> -			continue;
>> -		}
>> -
>> -		test_content_protection_on_output(output, s, content_type);
>> -		valid_tests++;
>> -	}
>> -
>> -	igt_require_f(valid_tests, "No connector found with HDCP capability\n");
>> -}
>> -
>>   static void test_content_protection_cleanup(void)
>>   {
>>   	igt_display_t *display = &data.display;
>> @@ -651,58 +616,63 @@ static void test_content_protection_cleanup(void)
>>   	}
>>   }
>>   
>> -igt_main
>> -{
>> -	igt_fixture {
>> -		data.drm_fd = drm_open_driver_master(DRIVER_ANY);
>> +static bool
>> +is_valid_output(igt_output_t *output, int content_type) {
>> +	if (!output->props[IGT_CONNECTOR_CONTENT_PROTECTION])
>> +		return false;
>>   
>> -		igt_display_require(&data.display, data.drm_fd);
>> +	if (!output->props[IGT_CONNECTOR_HDCP_CONTENT_TYPE] && content_type)
>> +		return false;
>> +
>> +	if (content_type && !sink_hdcp2_capable(output)) {
>> +		igt_info("\tSkip %s (Sink has no HDCP2.2 support)\n",
>> +			 output->name);
>> +		return false;
>> +
>> +	} else if (!sink_hdcp_capable(output)) {
>> +		igt_info("\tSkip %s (Sink has no HDCP support)\n",
>> +			 output->name);
>> +		return false;
>>   	}
>> +	return true;
>> +}
>>   
>> -	igt_subtest("legacy") {
>> +static void
>> +test_content_protection_dynamic(igt_output_t *output, int content_type)
>> +{
>> +
>> +	igt_dynamic_f("legacy-%s", output->name) {
>>   		data.cp_tests = 0;
>> -		test_content_protection(COMMIT_LEGACY, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_LEGACY,
>> +						  content_type);
>>   	}
>>   
>> -	igt_subtest("atomic") {
>> +	igt_dynamic_f("atomic-%s", output->name) {
>>   		igt_require(data.display.is_atomic);
>>   		data.cp_tests = 0;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>>   	}
>>   
>> -	igt_subtest("atomic-dpms") {
>> +	igt_dynamic_f("atomic-dpms-%s", output->name) {
>>   		igt_require(data.display.is_atomic);
>>   		data.cp_tests = CP_DPMS;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>>   	}
>>   
>> -	igt_subtest("LIC") {
>> +	igt_dynamic_f("LIC-%s", output->name) {
>>   		igt_require(data.display.is_atomic);
>>   		data.cp_tests = CP_LIC;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>>   	}
>>   
>> -	igt_subtest("type1") {
>> -		igt_require(data.display.is_atomic);
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_1);
>> -	}
>> -
>> -	igt_subtest("mei_interface") {
>> -		igt_require(data.display.is_atomic);
>> -		data.cp_tests = CP_MEI_RELOAD;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_1);
>> -	}
>> -
>> -	igt_subtest("content_type_change") {
>> -		igt_require(data.display.is_atomic);
>> -		data.cp_tests = CP_TYPE_CHANGE;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_1);
>> -	}
>> -
>> -	igt_subtest("uevent") {
>> +	igt_dynamic_f("uevent-%s", output->name) {
>>   		igt_require(data.display.is_atomic);
>>   		data.cp_tests = CP_UEVENT;
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>>   	}
>>   
>>   	/*
>> @@ -713,7 +683,7 @@ igt_main
>>   	 *  either of these options, we test SRM writing from userspace and
>>   	 *  validation of the same at kernel. Something is better than nothing.
>>   	 */
>> -	igt_subtest("srm") {
>> +	igt_dynamic_f("srm-%s", output->name) {
>>   		bool ret;
>>   
>>   		igt_require(data.display.is_atomic);
>> @@ -721,7 +691,78 @@ igt_main
>>   		ret = write_srm_as_fw((const __u8 *)facsimile_srm,
>>   				      sizeof(facsimile_srm));
>>   		igt_assert_f(ret, "SRM update failed");
>> -		test_content_protection(COMMIT_ATOMIC, HDCP_CONTENT_TYPE_0);
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>> +	}
>> +
>> +	/*
>> +	 * The tests mei_interface and content_type_change are meant only
>> +	 * for HDCP2.2 protocol, hence ignore for HDCP1.4.
>> +	 */
>> +	if (content_type != HDCP_CONTENT_TYPE_1)
>> +		return;
>> +
>> +	igt_dynamic_f("mei_interface-%s", output->name) {
>> +		igt_require(data.display.is_atomic);
>> +		data.cp_tests = CP_MEI_RELOAD;
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>> +	}
>> +
>> +	igt_dynamic_f("content_type_change-%s", output->name) {
>> +		igt_require(data.display.is_atomic);
>> +		data.cp_tests = CP_TYPE_CHANGE;
>> +		test_content_protection_on_output(output, COMMIT_ATOMIC,
>> +						  content_type);
>> +	}
>> +}
>> +
>> +igt_main
>> +{
>> +	int i;
>> +	struct cp_protocol {
>> +		const char *name;
>> +		int content_type;
>> +	} protocol[] = {
>> +		{ "hdcp-1_4", HDCP_CONTENT_TYPE_0 },
>> +		{ "hdcp-2_2", HDCP_CONTENT_TYPE_1 } };
>> +
>> +	igt_fixture {
>> +		data.drm_fd = drm_open_driver_master(DRIVER_ANY);
>> +
>> +		igt_display_require(&data.display, data.drm_fd);
>> +	}
>> +
>> +	for (i = 0; i < 2; i++) {
>> +		bool force_mei_hdcp_off = false;
>> +
>> +		/* In case a panel supports both HDCP2.2 and HDCP1.4 protocols,
>> +		 * HDCP2.2 is always chosen as default by the driver, even for
>> +		 * the TYPE 0 content.
>> +		 *
>> +		 * To forcefully test HDCP1.4, mei_hdcp module needs to be
>> +		 * unloaded. This takes away the HDCP2.2 support from the
>> +		 * platform, and HDCP1.4 can be tested. After the test the
>> +		 * module is loaded again and HDCP2.2 support gets re-enabled.
>> +		 */
>> +		if (i == HDCP_14 && igt_kmod_is_loaded("mei_hdcp")) {
>> +			igt_kmod_unload("mei_hdcp", 0);
>> +			force_mei_hdcp_off = true;
>> +		}
>> +
>> +		igt_subtest_with_dynamic_f("%s", protocol[i].name) {
>> +			igt_output_t *output;
>> +			int content_type = protocol[i].content_type;
>> +
>> +			for_each_connected_output(&data.display, output) {
>> +				if (!is_valid_output(output, content_type))
>> +					continue;
>> +				test_content_protection_dynamic(output,
>> +								content_type);
>> +			}
>> +		}
>> +		if (force_mei_hdcp_off)
>> +			igt_kmod_load("mei_hdcp", NULL);
>>   	}
>>   
>>   	igt_fixture {
>>

_______________________________________________
igt-dev mailing list
igt-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/igt-dev

  reply	other threads:[~2020-03-23  5:04 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-20  9:03 [igt-dev] [PATCH i-g-t] tests/kms_content_protection: Bifurcate tests into HDCP1.4 & 2.2 Ankit Nautiyal
2020-03-20  9:58 ` [igt-dev] ✗ GitLab.Pipeline: failure for " Patchwork
2020-03-20 10:15 ` [igt-dev] ✗ Fi.CI.BAT: " Patchwork
2020-03-20 11:05   ` Petri Latvala
2020-03-20 13:51     ` Vudum, Lakshminarayana
2020-03-20 10:28 ` [igt-dev] [PATCH i-g-t] " Petri Latvala
2020-03-23  3:59   ` Nautiyal, Ankit K
2020-03-20 12:46 ` Ramalingam C
2020-03-23  4:01   ` Nautiyal, Ankit K
2020-03-20 13:43 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2020-03-20 16:03 ` [igt-dev] [PATCH i-g-t] " Peres, Martin
2020-03-23  5:04   ` Nautiyal, Ankit K [this message]
2020-03-23  5:22     ` Anshuman Gupta
2020-03-23  6:33   ` Ramalingam C
2020-03-20 17:07 ` [igt-dev] ✗ Fi.CI.IGT: failure for " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=aae888e6-b9fc-5ffb-57f5-645b7dad8b19@intel.com \
    --to=ankit.k.nautiyal@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=martin.peres@intel.com \
    --cc=petri.latvala@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.