HID stops working when an image is inserted

[mangelajo]

The kvm managed device is an intel N100 NUC with linux, video works correctly.

1) Boot up, connect 2) I can type on the linux console of the NUC. 3) Insert an iso image 4) I can't type on the linux console of the NUC. 5) Eject the iso image 6) I can type again.

sometimes Reset HID will help (instead of ejecting the ISO image), sometimes it won't.

[razuritta]

Yup it seems that there's issue with hid emulation when changing usb config like turning on and off virtual disk/ethernet and image mounting. You might be able to see when using watch lsusb -tv command to see which part is missing tho.

[modzilla99]

I am affected as well, the issue still persists on v2.0.7.

Eject the iso image I can type again.

That's something I actually cannot confirm...

EDIT: Now I do, did not disable the Virtual USB toggle, unmounting the iso did not work.

EDIT2: When I toggle the button first, it does not work. Doing it the way around does though

You might be able to see when using watch lsusb -tv command to see which part is missing tho.

The output does not change (Virtual Network=Off and Virtual USB=On) and still shows:

Bus 001 Device 001: ID 1d6b:0002

This is the kernel log when I attach an ISO, I truncated it a bit since it will spam the signalling skipped: is not allowed by host and usb_gadget_wakeup msg:

Sep 11 18:40:06 kvm-4a10 daemon.info init: Id "acm" respawning too fast: disabled for 5 minutes
Sep 11 18:40:12 kvm-4a10 kern.info kernel: [  124.876906] dwc2 4340000.usb: wakeup: signalling skipped: is not allowed by host
Sep 11 18:40:12 kvm-4a10 kern.err kernel: [  124.876922] configfs-gadget gadget: usb_gadget_wakeup
Sep 11 18:40:12 kvm-4a10 kern.info kernel: [  124.924029] dwc2 4340000.usb: wakeup: signalling skipped: is not allowed by host
Sep 11 18:40:12 kvm-4a10 kern.err kernel: [  124.924045] configfs-gadget gadget: usb_gadget_wakeup
...
Sep 11 18:40:15 kvm-4a10 kern.info kernel: [  127.700560] dwc2 4340000.usb: bound driver configfs-gadget
Sep 11 18:40:15 kvm-4a10 kern.info kernel: [  127.818114] dwc2 4340000.usb: new device is high-speed
Sep 11 18:40:16 kvm-4a10 kern.info kernel: [  127.992103] dwc2 4340000.usb: new address 20
Sep 11 18:40:16 kvm-4a10 kern.info kernel: [  128.001259] dwc2 4340000.usb: wakeup: signalling skipped: is not allowed by host
Sep 11 18:40:16 kvm-4a10 kern.err kernel: [  128.001276] configfs-gadget gadget: usb_gadget_wakeup
Sep 11 18:40:16 kvm-4a10 kern.info kernel: [  128.303250] dwc2 4340000.usb: wakeup: signalling skipped: is not allowed by host
Sep 11 18:40:16 kvm-4a10 kern.err kernel: [  128.303265] configfs-gadget gadget: usb_gadget_wakeup
Sep 11 18:40:16 kvm-4a10 kern.info kernel: [  128.303354] dwc2 4340000.usb: wakeup: signalling skipped: is not allowed by host
Sep 11 18:40:16 kvm-4a10 kern.err kernel: [  128.303365] configfs-gadget gadget: usb_gadget_wakeup
...

[SmartyKNX]

Any news on this topic? Using a KVM without the ability to mount an ISO image to reinstall a remote server is pretty useless to me.

[modzilla99]

I was going to mention that it is now fixed, but I didn't find the issue anymore.

[SmartyKNX]

I can't confirm that this has been fixed. Currently my nanoKVM is connected to an HP Prodesk and as soon as I mount an ISO image I can no longer use the keyboard.

[modzilla99]

Have you tried different USB ports, specifically USB 2.0 ones?

[SmartyKNX]

According to HP, the ProDesk 600 G4 Mini does not have any USB 2.0 ports, so I have to use one of the USB3.1-Ports

Maybe the ISO malfunction is related to the keyboard issue mentioned here

[SmartyKNX]

I'm curious. Is this a known bug that someone at Sipeed is trying to fix, or are my 5 devices some kind of expensive doorstop?

I mean, a remote KVM has to be completely reliable because the only reason for a KVM is to be able to use it without being at a remote location. If I have to go there because the KVM is unstable or not working at all, I don't see many use cases for a KVM like this.

[larryqiann]

I'm curious. Is this a known bug that someone at Sipeed is trying to fix, or are my 5 devices some kind of expensive doorstop?

I mean, a remote KVM has to be completely reliable because the only reason for a KVM is to be able to use it without being at a remote location. If I have to go there because the KVM is unstable or not working at all, I don't see many use cases for a KVM like this.

I'd second this. While it's impressive that they shipped H.264 support pretty quickly, I think it's much more important for the base functionality to work correctly before any addons like this or Tailscale etc.