i.MX6DQ HDMI cannot send CEC correctly.

[satoshishimoda]

Hi community,

I have a question about i.MX6SDL HDMI.

Our partner is testing HDMI CEC fucntion with mxc_cec_test in L3.0.35_4.1.0 unit_test, but getting CEC version does not work well.

I saw hdmi mxc_cec_test.out function is not working, I think this issue is same as it, but it has not been resolved yet.

The environment of this issue is as below.

=====

[Board]

MCIMX6DL-SDP

[BSP]

L3.0.35_4.1.0

libCEC in L3.0.35_4.1.0

HDMI CEC is enabled

[HDMI Display]

TV which supports HDMI CEC or

HDMI analyzer

=====

Actually, our partner confirmed the CEC_TX_DATA value after write a value (0x409F) that is operation code to get CEC version.

Then, CEC_TX_DATA kept the value they wrote (0x409F) certainly.

However, actual output data was different.

Would you give me your comment why HDMI CEC cannot send the correct data to get CEC version?

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

had customer replaced the i.MX6 part on the SABRE-SD with an HDCP enabled

(first of all make sure the chip have fuse with fsl hdcp key) ?

Full detailed steps are given on links below (please contact local FAE for getting access)

i.MX6Q HDCP Integration help for SmarTV (iMX_6_hdcp_patch_bundle app)

HDCP 1.x support in android BSP

Best regards

igor

[satoshishimoda]

Dear Igor,

Thank you for your reply.

> had customer replaced the i.MX6 part on the SABRE-SD with an HDCP enabled

No, they nave not yet.

Can they use HDMI CEC only with HDCP enabled chip?

Best Regards,

Satoshi Shimoda

[igorpadykov · Accepted solution]

Hi Satoshi

sorry, I modified previous answer.

Correct would be:

>Can they use HDMI CEC only with HDCP enabled chip?

HDMI CEC can be used without HDCP enabled chip.

Seems L3.0.35_4.1.0 mxc_cec_test does not work properly,

some customers succeded to work with it modifying driver, as below

utilite/mxc_hdmi-cec.c at 8de44c2f1d770e306f7b22636e9c0ad6937d86d4 · wolfgar/utilite · GitHub

Best regards

igor

[satoshishimoda]

Dear Igor,

Thank you for your reply.

Our partner tested the code in the above GitHub, but CEC library initialization error was occurred.

Then, we want to confirm some points about the GitHub code.

Q1.

Have Freescale tested the GitHub code with SABRE board ?

Q2.

Is the GitHub code released by Freescale?

Actually, our partner worries whether Freescale supports the GitHub code or not.

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

I believe git codes are supported by community, as described on below link

https://community.freescale.com/message/402940#402940

Best regards

igor

[satoshishimoda]

Dear Igor,

OK, I understood the github code is supported by community, not Freescale official.

I guess it is the root cause why CEC library initialization error was occurred.

According to the Linux release note (IMX6DLLXRNSSD) of L3.0.35_4.1.0,  "HDMI CEC support"in chapter 3.1.

So I guess HDMI CEC works fine with L3.0.35_4.1.0 official BSP and SABRE board without the github code, right?

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

yes, right it works, probably except cec unit tests.

Linux release note does not mention unit tests.

Best regards

igor

[satoshishimoda]

Dear Igor,

Thank you for your quick response.

To make sure, I want to confirm one more thing about the Linux release note.

"HDMI CEC support"

Is this mentioned about CEC via HDMI_DDCCEC pad? or via KEY_ROW2?

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

DDCCEC pad is dedicated ground,

KEY_ROW2 is signal itself.

BTW is it the same question as :

i.MX6SDL HDMI CEC and DCD.   ?

what is "DCD"  there ?

Best regards

igor

[satoshishimoda]

Dear Igor,

> DDCCEC pad is dedicated ground,

> KEY_ROW2 is signal itself.

So BSP supports HDMI_TX_CEC_LINE signal via KEY_ROW2, not supports HDMI_TX_DDC_CEC via HDMI_DDC_CEC pad, right?

Actually, I mentioned HDMI_TX_CEC_LINE signal via KEY_ROW2 as "CEC", not HDMI_TX_DDC_CEC via HDMI_DDC_CEC pad.

So I wanted to confirm whether we are talking about same signal.

If BSP supports HDMI_TX_CEC_LINE, I believe HDMI driver does not occur this problem.

So the root cause is mxc_cec_test or other problem our customer occurred.

> BTW is it the same question as :

> i.MX6SDL HDMI CEC and DCD.   ?

> what is "DCD"  there ?

Sorry, I mistook, it is "DDC", not "DCD".

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

>So BSP supports HDMI_TX_CEC_LINE signal via KEY_ROW2, not supports HDMI_TX_DDC_CEC via HDMI_DDC_CEC pad, right?

KEY_ROW2 is signal itself (HDMI_TX_CEC_LINE).

DDCCEC pad is return signal (ground).

customer can connect HDMI_TX_DDC_CEC to hdmi

external device with R653.

Best regards

igor

[satoshishimoda]

Dear Igor,

Sorry, I misunderstood.

> DDCCEC pad is return signal (ground).

I found HDMI_DDCCEC is "Analog Ground" pad in IMX6SDLCEC (Rev.4), so I understood we cannot use this pad as CEC TX/RX between HDMI display, right?

According to these information, Freescale has tested CEC on KEY_ROW2 pad, right?

If you can, would you let me know what HDMI display was used when Freescale tested CEC function?

Now, we are doubting the HDMI display as root cause, so we want to prepare the same test environment as you tested.

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

since HDMI_DDCCEC is "Analog Ground" pad it can not

output data. KEY_ROW2 pad can output data.

You are right: several displays should be tested, not one.

Best regards

igor

[satoshishimoda]

Dear Igor,

> several displays should be tested, not one.

Could you let me know the tested display vendor, and part number if you can?

(e.g. LG, Samsung, Sony, etc...)

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

sorry, I have not such info

Best regards

igor

[satoshishimoda]

Dear Igor,

OK, I got it.

By the way, I believe the chip revision and SABRE board's revision don't effect to the HDMI CEC behavior, right?

Our partner is using rev 1.1 chip, but I guess it is no problem.

If you have confirmed the CEC with rev 1.1 chip also, would you let me know it?

Best Regards,

Satoshi Shimoda

[igorpadykov]

Hi Satoshi

yes this revision will work fine.

If you are asking about validation HDMI CEC results,

I would suggest to request them through local FAE channel.

Also one can look at sect.3 Test results AN4671

Best regards

igor

[satoshishimoda]

Dear Igor,

OK, I got it.

BTW, I got some new information about CEC problem from our partner.

But I cannot post a part of the information to public space.

So I create a new SR (SR# 1-3744532874).

Let us talk about the CEC problem in the new SR.

Best Regards,

Satoshi Shimoda

[igorpadykov]

HI Satoshi

I would suggest to work with that SR with local FAE,

since this SR requires additional info from third party,

note this module is third party IP. All available

info is in Reference Manual.

Best regards

igor