AM5716: Errata i896

Hi Kadowaki,

Atsushi Kadowaki said:

Q1 Is option 2 already supported by SDK 3.0?

I checked the code only option 1 is there , is it not working for you ?

Regards 

Diwakar 

Diwakar-san,
Thank you for your reply. Let me continue with my question.

Q3 You mentioned that only option 1 is available, could you please refer us to the code you have identified?

We are running into an issue with our USB devices not re-detecting.
The version of the SDK we are using is 6.03.00.106 (Linux).
The log at the time is as follows

　usb 1-1: timeout: still 3 active urbs on EP #1
　usb 1-1: timeout: still 3 active urbs on EP #1
　usb 1-1: timeout: still 12 active urbs on EP #82
　usb 1-1: timeout: still 12 active urbs on EP #82
　xhci-hcd xhci-hcd.0.auto: xHCI host controller not responding, assume dead
　xhci-hcd xhci-hcd.0.auto: HC died; cleaning up
　xhci-hcd xhci-hcd.0.auto: xHCI host not responding to stop endpoint command.
　xhci-hcd xhci-hcd.0.auto: xHCI host controller not responding, assume dead
　xhci-hcd xhci-hcd.0.auto: HC died; cleaning up

Q4 Are these logs related to errata i896?

This is a re-question of Q2.
The PORTSC register and the word PP are not listed on the TRM. Does it not exist in AM5716?

Thank you in advance for your answer.
Best regards,
Kadowaki

Hi Kadowaki,

Atsushi Kadowaki said:

Q3 You mentioned that only option 1 is available, could you please refer us to the code you have identified?

in the thread https://e2e.ti.com/support/processors-group/processors/f/processors-forum/530492/am572x-errata-i896?tisearch=e2e-sitesearch

You can check these commits 

6be136b usb: xhci: add quirk flag for broken PED bits
b268526 usb: dwc3: core: define macros for newest revisions
01245fb usb: host: xhci: add broken pe quirk flag to pdata
6f0d90c usb: host: xhci-plat: enable BROKEN_PE quirk if platform requested
913b397 usb: dwc3: host: pass BROKEN_PE flag for known broken revisions

steps to get the code already there in above e2e link.

You can find the changes related to option 1 .

Atsushi Kadowaki said:

This is a re-question of Q2.
The PORTSC register and the word PP are not listed on the TRM. Does it not exist in AM5716?

Shreyas can comment on this.

Regards

Diwakar

Hi Diwakar, Shreyas,

Thank you for your customer’s supports.

I tried to summarize remaining inquires as follows.

1. The PORTSC register and the word PP are not listed on the TRM. Could you share more those registers information, please?
2. Customer encountered Re-detecting issue. The version of the SDK customer is using is 6.03.00.106 (Linux). The log at the time is as follows. Would you think the following log is related to “errata i896” ?

　usb 1-1: timeout: still 3 active urbs on EP #1

　usb 1-1: timeout: still 3 active urbs on EP #1

　usb 1-1: timeout: still 12 active urbs on EP #82

　usb 1-1: timeout: still 12 active urbs on EP #82

　xhci-hcd xhci-hcd.0.auto: xHCI host controller not responding, assume dead

　xhci-hcd xhci-hcd.0.auto: HC died; cleaning up

　xhci-hcd xhci-hcd.0.auto: xHCI host not responding to stop endpoint command.

　xhci-hcd xhci-hcd.0.auto: xHCI host controller not responding, assume dead

　xhci-hcd xhci-hcd.0.auto: HC died; cleaning up
　

I’m sorry for rushing you, but we really appreciate if you will share advice/comments on those inquires. Customer needs to fix this issue as soon as possible.

Best regards,

Miyazaki

Diwakar, Shreyas,

Can we have your expert's advice/comments on those inquires, please?

Best regards,

Miyazaki

Miyazaki,

We are still waiting on the response.

Shreyas,

Thank you for sharing current status. I'm hoping you to gain comments from your expert's team before Christmas vacation.

Best regards,

Miyazaki

Miyazaki-san,

Unfortunately, my contact person is out on vacation until the new year. I will try to reach out to others for help.

Miyazaki-san,

The response I have received:

"There are like 3 layers:

• Xchi spec defines the generic register
• The DWC controller defines an instance
• TI wraps (bus+clock+power) and integrates a couple instances of the DWC into our SOC

-1-

XHCI:

The PORTSC (port status and control register) and the PP[9] (port power) field are in the xchi spec.  Their offset is defined there.

• Address: Operational Base + (400h + (10h * (n–1)))

-2-

DWC:

The DWC controller figures strangely compress the field names:

-3-

TI:

-Thanks,

Shreyas

Shreyas-san,

Thank you for contacting us.

Thank you for revealing the PORTSC register.
We will check it.

By the way, we have not yet received the following answer,

2.Customer encountered Re-detecting issue. The version of the SDK customer is using is 6.03.00.106 (Linux). The log at the time is as follows. Would you think the following log is related to “errata i896” ?

We are waiting for your answer here as well.


Best regards,
Kadowaki

Diwakar might help with the other question.