Do you have by chance an old USB 2.0 only flash drive? I don’t think it’s dead (yet). The simple test would be buying an ubiquitious pl2303 USB to UART and trying it in every socket. It shall work with both revisions. It works for me (of course if you have driver compiled into Linux).
But if it’s truly defect, maybe you shall not give up to return. It’s your right as a customer. The only question is time tho.
I don’t think I have any USB 2.0 flash drives. I think I have that USB to UART gadget. Will it work with kernel in a standard build? And how will I use it for testing?
I decided to keep defective board but now I noticed something strange. On that board there are two processes that are constantly running and taking 50% of CPU each. One is kworker and it seems to have something to do with USB. ( it is marked “usb_hub_wq”). Second one is ksoftirqd. I don’t see anything like that on non-defective board, it’s truly idling. My wild guess something in USB generates continuous interrupts and it leads to continuous CPU load. Does that makes sense?
BTW SOC is running quite a bit hotter on touch on defective board.
Yes, it will work outright. UART debug output and console are on by default, so you only need to connect each RX with each TX, and GND with GND.
115200, 8N1.
Among the inexpensive ones, I recommend cp2104. The pl2303 is not a great chip, although it is still better than the cursed ch340 ones.
Thanks for suggestion. Apparently I got couple of these converters, one with PL2303 and another with FT232RL chip. I guess either one will do. But due to an issue with USB seemingly causing endless interrupts which causes constant significant load I’m returning this board and getting a replacement.
Interesting.
NVME has a higher priority than SD-card. Are you sure the board started from SD-card？
Therefore, it is recommended to unplug NVME and check the USB hardware with SD-card at first.
I started with booting from SD card. While NVME SSD was connected it did not even had and image on it so could not have been used for booting or running. The USB was not working.
Sure, plug SD-card and NVME, if SD-card has image and NVME has not image, My v
isionfive 2 can start from SD card normallly with Nor-Flash Booting mode. But when unplug and plug and my USB2.0 mouse , the printing information is normal by serial port.
unplug USB2.0 mouse, the information is as follows
[ 1403.276802] usb 1-1.2: USB disconnect, device number 9
plug USB2.0 mouse，the information is as follows.
[ 1484.736519] usb 1-1.2: new low-speed USB device number 11 using xhci_hcd
[ 1484.884004] input: USB Optical Wheel Mouse as /devices/platform/soc/2b000000.pcie/pci0000:00/0000:00:00.0/0000:01:00.0/usb1/1-1/1-1.2/1-1.2:1.0/0003:1BCF:0002.0009/input/input10
[ 1484.900563] hid-generic 0003:1BCF:0002.0009: input: USB HID v1.11 Mouse [USB Optical Wheel Mouse] on usb-0000:01:00.0-1.2/input0
Thanks for your suggestion, but with the help of folks on this forum I diagnosed the problem as being a hardware failure and I’m returning this board.
As you mentioned, you connnected USB 3.0 external SSD and it works. It looks like the hardware is fine. It is strange that only the USB2.0 mouse and keyboard were affected。I wonder if you have all four USB3.0 ports with external devices at the same time。 Otherwise， I wonder if what’s the supply voltage and current of the board.
According to @mzs you can have a hardware fault which only affects USB 2.0 and not 3.0. It appears to be my case. I only connected simple wired keyboard, not several devices and I use power brick which provides 68W. I since got two more boards and they both fine.
I am having the same USB issue, so this seems not uncommon.
However, as I can use the machine without USB (as a server), I’ll be keeping it.