Patch to make i.MX6DQ USB to support test modes for certification test

[#1]

According to EHSET_v1.01.pdf description (http://testusb.com/EHHS.html), the USB physical certification test are from 2 sides. One test case is PC running USB-IF HSET utility as being USB host to test the target board being USB device; the other case is target board acting as USB host to initate test signals to a test fixture.

The principle of test modes supported are based on Vid/Pid pair  during enumration by host whoever takes.

Apply the attached patch (I use r13.4.1 kernel_imx tree), then enable "FSL_USB_TEST_MODE" in kernel config, rebuild uImage to get the test modes support kernel.

[satoshishimoda]

Hi Xia Zhang,

Could you let me know whether the patch has been applied to newer BSP?

(e.g. L3.10.17_1.0.0-ga, kk4.4.2_1.0.0-ga, etc...)

Best Regards,

Satoshi Shimoda

[XiaoPong]

Hi, Shimoda San,

     The patch was applied to KK4.4.2_1.0.0-GA.

Best Regards,

Wilson

[satoshishimoda]

Hi Wilson-san,

Thank you for your reply.

Then, do you know how about Linux BSP?

Especially, I want to know whether it has been applied to L3.10.17_1.0.0-ga (Yocto BSO).

Best Regards,

Satoshi Shimoda

[XiaoPong]

Good morning, Shimoda San,

     Linux does not need any patch but there are commands directly changing register content in order to generate USB TEST pattern and they work on L3.0.35_4.1.0:

            1 ./memtool -32 0x020c4080 1  (to read current value; for example, 0xc00)

            2 ./memtool -32 0x020c4080=(current value + 0x3; for example, 0xc03)

            3 ./memtool -32 0x02184140=0x2

4 ./memtool -32 0x021841A8=0x3

5 ./memtool -32 0x02184184=0x1C040000  ( * for some reason, first assignment does not generate test pattern. In this case, redo this command.)

     However, in my case running on Yocto, the 5th command changes bit0 and 1 of 0x020C4080 from 1 back to 0. So, we have to run 2nd command to set clock once more; otherwise, any following command could cause system crash. While I have submitted this issue to our internal team for help, you may try these commands on L3.0.35_4.1.0.

Best Regards,

Wilson

[宗標廖]

Hello, Wilson

I am now working on i.MX6 with L3.0.35_4.1.0 and I want to test the High Speed eye pattern(for certification test) when i.MX6 working on device mode.I don't know whether the following operations are right? Could you help me?

1)./memtool -32 0x020c4080 1

2)./memtool -32 0x202c4080=(current value + 0x3)

3)./memtool -32 0x02184140=0x2

4)./memtool -32 0x021841A8=0x2 (set to Device Controller)

5)./memtool -32 0x02184184=0x1C050000 (According to IMX6SDLRM.pdf, I set PTC field to FORCE_ENABLE_HS to enable High Speed).

Could you tell me if is this right?

Thank you!

Best Regards,

ZongbiaoLiao

[XiaoPong]

Hi, Zongbiao,

     Haven't set 0x2 into 0x021841A8. After my step 1~5, test pattern should be generated.

Good Luck!

[宗標廖]

Hi,Wilson

Thank you for replying me so quickly.

In my case,  i.MX6  is working on USB OTG device mode.

if I run ./memtool -32 0x021841A8=0x3, i.MX6 will be work on host mode.

So I think I need to set 0x021841A8 to 0x2, is it right?

Best Regards,

ZongbiaoLiao

[XiaoPong]

Hi, Zongbiao,

     This article is talking about HOST mode, not DEVICE mode. Being device mode, no change is required.

Best Regards,

Wilson

[marcowang]

Hi All,

Is it work at Android L5.1.2 source ? Does someone test it ?

Thanks for your help !!

BR

[huanghsiu-ling]

Hi,

How can I set iMx6dl USB into test mode (host) on kernel 4.1.15?

Above register setting dose not work on kernel 4.1.15

[caocai]

Hi  Xia Zhang,

   I take the patch in kernel3.10.53. But the kernel  print "in test mode, but single step NOT on" all the way.

 Could you give some  advice?

Best Regards,

George