USB3.0 incompatible with USB1.1

[dr_berta]

Hi all,
I have a problem with a new panel pc. It has 4 USB3.0 ports usually driven with the kernel driver xhci.
Now I have to connect to the pc a device with USB1.1. When I do it it works bad and fails all the reboot with the following error (dmesg):

"usb1-1: device descriptor read/64, error -71

xhci_hcd 0000:00:14.0: Setup ERROR: setup context command for slot 2

usb1-1: Device not responding to setup address

usb 1-1: device not accepting address 5, error -71

hub 1-0:1.0: unable to enumerate USB device on port 1"

After that the system disables all the usb ports

I tried to downgrade the ports to usb 2.0 from the BIOS, but using the driver xhci I have always the same behaviour. I tried to eliminate the xhci driver using only the ehci driver, but in this case it is not possible to see the all the usb ports on the system.

Please, do you have any suggestion about how to solve this issue?

Thanks
Claudio

[NeddySeagoon]

dr_berta,

Please post the output of

[dr_berta]

Hi,
this is the result of the first command you asked:

[dr_berta]

Just an update.
I upgraded the kernel from 3.14.14 to 3.19.8 and the situation now is a little bit better.
Now at the reboot the error is shown in the screenshot dmesg.png you can find here:https://www.dropbox.com/sh/4dx944ew2hy4pen/AAA_1FfbYoGKCoVSXSBFcPDJa?dl=0

The dmesg2.png is the answer of the system when I tried to switch off and on the usb port with the unbind/bind command.

The dmesg3.png shows the result I have disconnecting and reconnecting the usb device

I sent the same info to the manufacturer of the panel PC and they answered me this:

[NeddySeagoon]

dr_berta,

You posted

[dr_berta]

Sorry for the wrong command.

Here you find the correct command:

[NeddySeagoon]

dr_berta,

[dr_berta]

Hi,
thanks for your info they are very useful.
I think the problem is due to the real structure of the usb into the pc. I try to describe it:

The panel pc has only 1 usb3.0 hub with 16 ports:
Only 4 ports are accessible for the user but only 3 are used

One port for the mouse

One port for the keyboard

One port for the usb isolator controlled by two hub: ID 0451:2046 Texas Instruments, Inc. TUSB2046 Hub.
This isolator uses an optical fiber for its purpose and has one hub before and one hub after the fiber.
After the fiber it is active, so it has a power supply to provide the nergy for the hub.

The usb isolator is connected to a 4 port active usb hub: ID 05e3:0605 Genesys Logic, Inc. USB 2.0 Hub

The two fpga boards are connected to this hub.
Since the hub is active, its power supply provides the current for the two fpga boards

From the command "lsusb -vvv" you can also see that the first hub of the usb isolator does not detect the presence of the second hub after the fiber, but only the device that is connected to it.
In this way the two fpga boards have no impact on the power of the usb hub of the pc.
So I'm not expecting to have a real consumption problem.

For this reason I'm thinking that the issue could be due to a communication problem between the hub of the usb isolator and the usb 3.0 hub.
Dmeag tipically is giving me this error:

[NeddySeagoon]

dr_berta,

I was expecting to tease your USB bus structure out of

[dr_berta]

NeddySeagoon,

I tried to do all the test you suggested.
The problem is present on all the systems having that model of pc
The problem is present only if I connect the fiber isolator (usb1.1) on the port used for it.

I tried to put a usb2.0 hub between the fiber isolator and the usb3.0 port of the pc. This eliminates the problem of the error (r=-71) at the reboot. In this condition, randomly and only on some systems, the usb communication with the fpga board fails with error r=-32 (an overcurrent?).

About your answer, I agree that it is not possible to derive the structure of the usb connections from the command

[NeddySeagoon]

dr_berta,

Overcurrent detection behaviours vary wildly. I don't know of any that add up the Maxpower devices declare.
Some devices do nothing ... the current can increase until something melts.
Some devices have a auto resetting fuse, either per port or per root hub. Its not really a fuse, more like a thermister. When it gets hot, it goes high resistance and blocks the current flow, then it cools and the cycle repeats.
Other devices do 'load shedding' some ports operate normally, others are shut down to keep the total current in limits.

The data sheet for your TUSB2046 Hub can be found from that link.
The upstream end is USB2 the downstream end is 4 USB1 ports.
Overcurrent detection is done here. The datasheet says that overcurrent detection is per port but control can be either ganged (all ports go off) or per per port.

Optoisolation is not mentioned but that's an implementation detail. The link is for the TI USB chip itself.

If you swap the two FPGA boards over, does the problem swap ports too?
i.e. does it follow the FPGA or stay with the the port?
_________________
Regards,

NeddySeagoon

Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail.

[dr_berta]

Hi,
I tried to swap the boards and the problem stays with the port. Better, I noted that the first device identified by the driver works fine, the other has the problem.

Anyway I found a way to solve the problem.
The solution I found is to replace the fiber cable with a new model usb3.0 certified and put after that a usb 3.0 hub to connect the two fpga boards.

With this chain everything works fine.

Thanks a lot for your help
Claudio