AM625: AM625 USB0 can only enumerate once as device

wang robin

Part Number: AM625
Other Parts Discussed in Thread: TPS65988

Hi, USB Owner;

I encountered a problem. After the system is loaded, set USB0 to peripheral mode. the device can only be enumerated once, the USB port cannot be enumerated again.

steps as follows:

insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/libcomposite.ko

insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/function/usb_f_printer.ko

insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/legacy/g_printer.ko

When the device cannot be enumerated, it needs to be powered off and then loaded to enumerate;

dv1912_usb_host.txt

25:39  [ 2461.258907]  [0:    kworker/0:0: 4780] usb 1-1.2: new high-speed USB device number 9 using xhci-hcd
25:39  [ 2461.376557]  [0:    kworker/0:0: 4780] usb 1-1.2: generic_probe: slot_id = 2, idVendor=0525, idProduct=a4a8
25:39  [ 2461.406625]  [0:    kworker/0:0: 4780] usblp 1-1.2:1.0: usblp0: USB Bidirectional printer dev 9 if 0 alt 0 proto 2 vid 0x0525 pid 0xA4A8
25:39  [ 2461.419147]  [0:    kworker/0:0: 4780] usb_notify: call_device_notify device
25:39  [ 2461.426358]  [0:    kworker/0:0: 4780] usb_notify: send_otg_notify event=device_connect(20) enable=1
25:39  [ 2461.435615]  [0:    kworker/0:0: 4780] usb_notify: extra_notifier_callback event=device_connect(20)
25:39  [ 2461.445244]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state+ event=device_connect(20), enable=on
25:39  [ 2461.454999]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state- event=device_connect(20), cable=none
25:39  [ 2461.464895]  [0:    kworker/0:0: 4780] usb_notify: check_gamepad_device : product=Printer Gadget
25:39  [ 2461.473815]  [0:    kworker/0:0: 4780] usb_notify: check_lanhub_device : product=Printer Gadget
25:39  [ 2461.482693]  [0:    kworker/0:0: 4780] usb_notify: call_battery_notify : VID : 0x525, PID : 0xa4a8, on=1, count=1
25:39  [ 2461.493516]  [0:    kworker/0:0: 4780] usb_notify: send_otg_notify event=smtd_ext_current(18) enable=1
25:39  [ 2461.502946]  [0:    kworker/0:0: 4780] usb_notify: extra_notifier_callback event=smtd_ext_current(18)
25:39  [ 2461.512318]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state+ event=smtd_ext_current(18), enable=on
25:39  [ 2461.522237]  [0:    kworker/0:0: 4780] usb_notify: No smart dock!!!!!!
25:39  [ 2461.528860]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state- event=smtd_ext_current(18), cable=none
25:39  [ 2461.539305]  [0:    kworker/0:0: 4780] usb_notify: check_device_speed : attached highspeed device
25:44  
25:45  dv1912:/ # 
25:45  dv1912:/ # 
25:45  dv1912:/ # 
25:45  dv1912:/ # 
25:45  dv1912:/ # 
25:45  dv1912:/ # 
25:45  dv1912:/ # 
26:09  dv1912:/ # [ 2490.989459]  [0:    kworker/0:0: 4780] usb 1-1.2: USB disconnect, device number 9
26:09  [ 2491.000492]  [0:    kworker/0:0: 4780] usblp0: removed
26:09  [ 2491.008565]  [0:    kworker/0:0: 4780] usb 1-1.2: generic_disconnect: slot_id = 2, idVendor=0525, idProduct=a4a8
26:09  [ 2491.019360]  [0:    kworker/0:0: 4780] usb_notify: call_battery_notify : VID : 0x525, PID : 0xa4a8, on=0, count=0
26:09  [ 2491.029749]  [0:    kworker/0:0: 4780] usb_notify: send_otg_notify event=smtd_ext_current(18) enable=0
26:09  [ 2491.039149]  [0:    kworker/0:0: 4780] usb_notify: extra_notifier_callback event=smtd_ext_current(18)
26:09  [ 2491.048513]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state+ event=smtd_ext_current(18), enable=off
26:09  [ 2491.058544]  [0:    kworker/0:0: 4780] usb_notify: No smart dock!!!!!!
26:09  [ 2491.065218]  [0:    kworker/0:0: 4780] usb_notify: extra_notify_state- event=smtd_ext_current(18), cable=none
26:14  
26:14  dv1912:/ # 
26:15  dv1912:/ #

am625_usb_device.txt

20:24:38
20:24:38 _____                    _____           _         _   
20:24:38|  _  |___ ___ ___ ___   |  _  |___ ___  |_|___ ___| |_ 
20:24:38|     |  _| .'| . | . |  |   __|  _| . | | | -_|  _|  _|
20:24:38|__|__|_| |__,|_  |___|  |__|  |_| |___|_| |___|___|_|  
20:24:38              |___|                    |___|            
20:24:38
20:24:38Arago Project http://arago-project.org am62xx-evm ttyS2
20:24:38
20:24:38Arago 2021.09 am62xx-evm ttyS2
20:24:38
20:24:40am62xx-evm login: 
20:24:42am62xx-evm login: root
20:24:43root@am62xx-evm:~# 
20:24:43root@am62xx-evm:~# 
20:24:43root@am62xx-evm:~# 
20:24:43root@am62xx-evm:~# 
20:24:57root@am62xx-evm:~# [   33.862246] tlv71033: disabling
20:25:34
20:25:34root@am62xx-evm:~# 
20:25:34root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:35root@am62xx-evm:~# 
20:25:38root@am62xx-evm:~# insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/libcomposite.ko
20:25:38insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/function/usb_f_printer.ko
20:25:38insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/legacy/g_printer.koroot@am62xx-evm:~# insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/function/usb_f_printer.ko
20:25:39root@am62xx-evm:~# insmod /lib/modules/5.10.109-g9e58028f94/kernel/drivers/usb/gadget/legacy/g_printer.ko
20:25:39[   75.463975] printer gadget: printer ready
20:25:39root@am62xx-evm:~# [   75.823567] printer gadget: Using interface 0
20:25:39[   75.832122] printer gadget: Using interface 0
20:25:40
20:25:40root@am62xx-evm:~# 
20:25:41root@am62xx-evm:~# 
20:26:05root@am62xx-evm:~# 
20:26:32root@am62xx-evm:~# [  128.546087] Initializing XFRM netlink socket

over 2 years ago

0 wang robin over 2 years ago

Prodigy 40 points

processor SDK：linux 08.03.00.005;

kernel version: 5.10.109

0 Andreas Dannenberg over 2 years ago in reply to wang robin

TI__Guru 62027 points

Hi! The subject owner is out of the office today. Please allow some time for a response. Thanks.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

Does the issue happen on your custom board or AM625 SK EVM?

If this is on your custom board, do you see the same issue on the SK EVM?

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

Hi Bin;

It has no problem on the SK EVM;

different from the SK EVM, there is no tps65988 on our board.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

I don't think the PMIC would be related.

Can you please share the USB portion of your board schematics showing the USB connections between AM625 and the USB receptacle?

Do you use Linux or Windows USB host? Have you checked the kernel log on the Linux host to see if the host reports anything after the first enumeration?

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

I received your board schematics offline.

The schematics shows both USB0_VBUS and USB1_VBUS pins are routed to the USB typeB connector (J9) pin1 (VBUS_IN). both pins shouldn't connect together. USB1_VBUS pin should go to the USB typeA connector (J7) pin1 (VBUS_5V0_TYPEA1).

Please correct this on your board and let me know if this solves the issue.

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

DC2221_MAIN_T2.pdf

This is our new schematic, help to check it please.

the problem remains

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

I have tested both Linux and windows USB Host. The am62x development board has the following log for each plug in and out, but our board only has them for the first time；

“printer gadget: Using interface 0”

After the first plug, there is no log on the host side.

0 wang robin over 2 years ago

Prodigy 40 points

Today, we did several experiments on our board:

1. Remove the 5V combiner diode

2. Remove the zener diode on VBUS

The IO level of USB0_VBUS is consistent with that on the development board, but there are still problems on our board.

then I deleted tps6598x.ko from the development board, and it can still be enumerated normally on the development board

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

The new board schematics seems to be fine on the USB design, but I am not sure if the resisters on the VBUS line R648 & R649 are too big. What is the USB0_VBUS pin voltage (near AM62x) after unplugged from the USB host? Is it discharged quickly to 0v?

Does the linux host dmesg log show any new messages when plugged in the second time?

Can you please probe the USB0_VBUS, USB0_DM, USB0_DP lines when plug in the second time using a scope?

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

When USB0_DP is 0.068v, the device will not be enumerated, and there is no new messages on linux host when plug-in USB cable.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

You initially mentioned the USB device can only be enumerated ONCE, but this new table shows the device has been enumerated twice. Is the number of successful enumeration always once or twice? or is it random?

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

We tested another board. The enumeration is random. It may be enumerated several times(the test experiment can enumerate up to five times), but after 0.068v, the device can't enumerate any more。

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

Instead of g_printer.ko, can you please test with g_zero.ko to see if it makes any difference?

0 Tony Tang over 2 years ago in reply to wang robin

TI__Mastermind 27191 points

The GND connection is different with AM62-SK. on customer board, connected shield GND with signal GND together, this may cause ESD issue.

on AM62-SK.

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

test with g_zero.ko, the problem remains.

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

Hi，Bin；

We also found the problem on AM625 SK EVM. The steps are as follows:

1.Power on AM625-SK, execute command：modprobe libcomposite; modprobe g_zero

2.Quickly plug and unplug the USB cable repeatedly, it will appear in about ten minutes.

After the problem occurred, warm reboot system and then insmod ko, the AM625 SK EVM cannot be enumerated, the Serial output “dwc3 31000000.usb: failed to enable ep0out”, can only be enumerated by power cycle the board.

If modify AM62-SK as below, it is more easy to reproduce the failure.

0 wang robin over 2 years ago in reply to Tony Tang

Prodigy 40 points

Add 1.2k pull-down resistor to VBUS on our board，Quickly plug and unplug the USB cable 10 minute，the problem did not recur.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

Quickly plug and unplug until the gadget is fully enumerated is not a normal use case, though I understand the USB shouldn't lockup regardless.

Back to the original issue on your board, when you initially observed the issue, how soon did you unplug the board from USB host? When the issue happened, did you see "dwc3 31000000.usb: failed to enable ep0out" on the board console too?

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

normal plugging will cause issues on our board, when the issue happened, the console has no message.

after excecute reboot command, then "dwc3 31000000.usb: failed to enable ep0out" on the board console.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

wang robin said:
after excecute reboot command, then "dwc3 31000000.usb: failed to enable ep0out" on the board console.

after reboot command, does this message got printed before or after the board is rebooted?

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

the message "dwc3 31000000.usb: failed to enable ep0out" got printed after the board is rebooted.

Add 1.2k pull-down resistor to VBUS on our board，we did the test item "quickly plug and unplug", the problem doesn't occur for now. It seems that the problem has been solved.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

wang robin said:
the message "dwc3 31000000.usb: failed to enable ep0out" got printed after the board is rebooted.

Is the AM62x board connected to the USB host while rebooting? I don't understand why this message would happen during reboot, or how it is related to this enumeration issue. Can you please attach the reboot log when this message got printed?

wang robin said:
Add 1.2k pull-down resistor to VBUS on our board，we did the test item "quickly plug and unplug", the problem doesn't occur for now. It seems that the problem has been solved.

I am looping in our USB HW expert to comment on this VBUS pull-down resister.

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

We found that this issue is related to the falling edge of USB0_VBUS when plug removed from the board.

The falling edge（70%~30% 1.8V） of USB0_VBUS lasts 8ms on EVM. When the falling edge lasts 80ms, this issue will occur easily. When occurs, it seems about USB IP core will be hang up with nothing console log.

Then we input "reboot" in console to reboot the board. After reboot finished, input "root", "modprobe libcomposite", "modprobe g_zero" , the message "dwc3 31000000.usb: failed to enable ep0out" will be printed.

The USB IP core cannot escape from the state of hanging up through hot reboot.

0 Bin Liu over 2 years ago in reply to wang robin

TI__Guru*** 147816 points

Hi Robin,

Thanks for the update.

I am routing this thread to our USB hardware expert for comments.

0 Tony Tang over 2 years ago in reply to Bin Liu

TI__Mastermind 27191 points

Bin，

If USB module is locked, is there anyway to check and recover?

0 Rajashri Hazarika over 2 years ago in reply to Bin Liu

TI__Intellectual 1515 points

Hello Bin,

Thank you for the query.

Can you please open a new thread for the hardware query for clarity.

Add 1.2k pull-down resistor to VBUS on our board, we did the test item "quickly plug and unplug", the problem doesn't occur for now. It seems that the problem has been solved.

I will review the thread and check on VBUS pull-down resister

Regards

Rajashri

0 Bin Liu over 2 years ago in reply to Tony Tang

TI__Guru*** 147816 points

Tony,

Tony Tang said:
If USB module is locked, is there anyway to check and recover?

It appears warm reset (Linux reboot command) doesn't fix it, I don't think there is any other way to recover it other than power cycle the board.

You can try to unload and reload the USB driver modules to see if that can recover it.

0 wang robin over 2 years ago in reply to Bin Liu

Prodigy 40 points

"You can try to unload and reload the USB driver modules to see if that can recover it."

when the issue happened, unload and reload the usb driver module can not work, it must be power off the board.

0 Tony Tang over 2 years ago in reply to Bin Liu

TI__Mastermind 27191 points

in order to reproduce the failure, customer add a 4.7uF capacitor parallel with R156 on USB0_VBUS on AM62-SK as below:

it extends the USB0_VBUS falling edge from 8ms to 30ms, plug and insert 10 times can reproduce the issue.

Reboot (warm reset) can't resume USB0 to work.

+1 Carey Wang over 2 years ago in reply to Bin Liu

Prodigy 30 points

Bin, please try to reproduce this problem on AM62-SK and solve this problem. We have tried many times for a long time before finding that the slow falling edge of USB0_VBUS can cause USB IP core stops working even after warm reboot. It is not s specific problem on our board.

0 Bin Liu over 2 years ago in reply to Carey Wang

TI__Guru*** 147816 points

Hi Carey,

I have routed this thread to our USB HW expert to look at the USB VBUS discharge circuit design.

0 -DK- over 2 years ago in reply to Carey Wang

TI__Mastermind 27890 points

Hi Carey,

Please review the capacitance on your VBUS trace. From the specification (Section 7.2.4.1) (emphasis mine):

The maximum load (CRPB) that can be placed at the downstream end of a cable is 10 µF in parallel with 44 Ω. The 10 µF capacitance represents any bypass capacitor directly connected across the VBUS lines in the function plus any capacitive effects visible through the regulator in the device. The 44 Ω resistance represents one unit load of current drawn by the device during connect.

We typically recommend starting values of ~4.7uF for new board designs. This may need to be tuned depending on board performance.

0 Kallikuppa Sreenivasa over 2 years ago in reply to -DK-

TI__Guru**** 158619 points

Hello Carey,

Can you please request customer to remove Q94 for testing.

Q94 will isolate the VBUS_IN from VBUS which has the major caps

Power should be applied through J12

Regards,

Sreenivasa

0 Kallikuppa Sreenivasa over 2 years ago in reply to Kallikuppa Sreenivasa

TI__Guru**** 158619 points

Hello Carey,

Checking if any further tests were done and customer has some resolution

Regards,

Sreenivasa

0 Kallikuppa Sreenivasa over 2 years ago in reply to Kallikuppa Sreenivasa

TI__Guru**** 158619 points

Hello Carey,

I am closing the thread since i have not heard from you.

Regards,

Sreenivasa

Because of the holidays, TI E2E™ design support forum responses will be delayed from Dec. 25 through Jan. 2. Thank you for your patience.

Processors

Processors forum

AM625: AM625 USB0 can only enumerate once as device