PROCESSOR-SDK-AM335X: USB error

Fariya Fatima

Hi,

Our intention was to disable ADP OTG pulsing which happens every 2 seconds and instead when the user presses a key, enable the VBUS. This was implemented using the patches from the above thread. One peculiar scenario we face is as follows:

Steps followed are as below

a) Plug in USB flash drive onto our unit

b) Give the command, echo 1 > /sys/kernel/debug/musb-hdrc.0/softconnect

c) Mount the USB and follow it up with USB file copy etc

d) Run a script which unmounts the USB and gives the command mentioned below to disconnect the USB flash drive

# echo 0 > /sys/kernel/debug/musb-hdrc.0/softconnect

e) At times after we run this command, we see the below logs:

usb usb1-port1: Cannot enable. Maybe the USB cable is bad?

If after doing step $d (with the flash drive still connected), we were to follow from step b onwards, the flash drive doesn't get mounted properly. Below is the log we get:

usb 1-1: USB disconnect, device number 41
usb 1-1: new high-speed USB device number 42 using musb-hdrc
usb-storage 1-1:1.0: USB Mass Storage device detected

A typical detection log is as below:

[ 9805.649560] usb 1-1: new high-speed USB device number 41 using musb-hdrc
[ 9805.827479] usb-storage 1-1:1.0: USB Mass Storage device detected
[ 9805.843293] scsi host0: usb-storage 1-1:1.0
[ 9807.269776] sd 0:0:0:0: [sda] 30274560 512-byte logical blocks: (15.5 GB/14.4 GiB)

[ 9807.270630] sd 0:0:0:0: [sda] Write Protect is off

[ 9807.270691] sd 0:0:0:0: [sda] Mode Sense: 45 00 00 00
[ 9807.271091] sd 0:0:0:0: [sda] Write cache: disabled, read cache: enabled, doesn't support DPO or FUA
[ 9807.284501] sda: sda1

F) If instead of immediately following up with step b after step d, we were to wait about 5-8 seconds then the flash drive gets detected properly. It appears that the logs shows the below line about 4 - 5 times and then the device gets disconnected:

[ 9865.519665] usb usb1-port1: Cannot enable. Maybe the USB cable is bad?
[ 9869.839760] usb usb1-port1: Cannot enable. Maybe the USB cable is bad?
[ 9874.159678] usb usb1-port1: Cannot enable. Maybe the USB cable is bad?
[ 9878.479809] usb usb1-port1: Cannot enable. Maybe the USB cable is bad?

[ 9878.480145] usb 1-1: USB disconnect, device number 43

After getting the disconnect device number.." log, we were to do enable the softconnect command, things work as expected. My question is why do I get the line "Cannot enable. May be the USB cable is bad" How do I overcome the delay before it gives the "USB disconnect, device number.." log.

Regards,

Fariya

over 6 years ago

0 Biser Gatchev-XID over 6 years ago

TI__Guru**** 393215 points

Please post which Linux version you are using.

0 Fariya Fatima over 6 years ago in reply to Biser Gatchev-XID

Intellectual 810 points

Kernel version 4.9.31.

0 Bin Liu over 6 years ago in reply to Fariya Fatima

TI__Guru**** 172861 points

Hi Fariya,

I think the issue is that the softconnect entry was not design for this scenario, which causes the usb driver stack confused.

What is the USB0 dr_mode setting in dts? I am going to replicate the scenario on my setup to understand what is happening.

0 Bin Liu over 6 years ago in reply to Fariya Fatima

TI__Guru**** 172861 points

Is your USB0 designed to be dual role mode or host-only mode?

0 Fariya Fatima over 6 years ago in reply to Bin Liu

Intellectual 810 points

Hi Bin Liu,

The dr_mode setting in dts is "otg".

Regards,

Fariya

0 Fariya Fatima over 6 years ago in reply to Bin Liu

Intellectual 810 points

USB0 is designed to be dual role mode but with the host mode enabled on-demand.

Regards,

Fariya

0 Fariya Fatima over 6 years ago in reply to Fariya Fatima

Intellectual 810 points

Hi Bin Liu,

Were you able to try for this issue on your setup?

Thanks

Fariya

0 Bin Liu over 6 years ago in reply to Fariya Fatima

TI__Guru**** 172861 points

Hi,

Fariya,

Sorry, I am in the middle of another work and haven't got a chance for your case yet. But the idea I have is that you don't need to use the softconnect sysfs entry when disconnect the usb device, you probably only need it to initiate the enumeration when a usb device is attached. I will try this on my evm and provide you a solution, I will probably look at this later this week. Sorry for the delay.

0 Fariya Fatima over 6 years ago in reply to Bin Liu

Intellectual 810 points

Hi Bin,

Thanks for the reply. We do need the

echo 0 > /sys/kernel/debug/musb-hdrc.0/softconnect

command given above as we intend to disconnect the connected flash drive if the user doesn't use it for about 30 secs of having mounted it.

Do help us with your observation after having tried the scenario I mentioned.

Regards,

Fariya

0 Fariya Fatima over 6 years ago in reply to Fariya Fatima

Intellectual 810 points

Hi Bin,

Were you able to have a look into the scenario I mentioned?

Thanks

Fariya

0 Fariya Fatima over 6 years ago in reply to Fariya Fatima

Intellectual 810 points

Hi Bin,

Did you have a chance to look into this?

Regards,

Fariya

0 Bin Liu over 6 years ago in reply to Fariya Fatima

TI__Guru**** 172861 points

Hi Fariya,

I apologize for the delay. I tested your scenario but was not able to reproduce this "Cannot enable. Maybe the USB cable is bad?" error. To align on what we have changed in the kernel, attached below are the 5 patches I have in SDK4.0.0.4/ti2017.01 kernel, which is based on v4.9.28, close to yours v4.9.31.

https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/791/softconnect_2D00_ti2017.01.tgz

Then I used the following sequence to test your scenario, I skipped step d) as I don't think it is relevant. I tested 50+ times but didn't see any error.

  #!/bin/sh                                                                       
  cd /sys/kernel/debug/musb-hdrc.0                                                
  echo 1 > softconnect                                                            
  sleep 10                                                                        
  echo 0 > softconnect                                                            
  sleep 5

Processors

Processors forum

PROCESSOR-SDK-AM335X: USB error