Advanced Search

Reading something today. Maybe try using smp affinity to lock the USB interrupt to  Core1 (away from core 0 housekeeping) and assuming > 2 cores. You can find the interrupt by
cat /proc/interrupts | grep -i usb

Then follow one of the many posts about cpu affinity for networking. Just remember you need to reset this on every boot.

Okay, so maybe we can conclude that somehow the hardware is not 100% suitable on your "real" setup?

If you still test on the N100, I believe you could force the setupAsyncTransfer error by putting the loops back to original values (or even higher), and jogging while a program is running.

Duplicated the setup to a N100 mini-PC.
Couldn't crash the xhc, didn't try for that long though.
So I unplugged the dongle. Put it back.
And it restarted and continued to work. Like magic.

It didn't fail at the usual setupAsyncTransfer, it was somewhere else.
 

Signals utilized for operating spindle and feed override. The feed rotary button can serve in

• Continuous move x% from max velocity
• Step move x mm
• MPG override feed/spindle
• The special position Lead.
  Continuous: In this mode jogging is performed at the selected feed rate. As long the jogwheel turns, the selected axis moves.
  Step: In this mode the machine moves steps  wheel_counts at the currently selected step size and the current set feed rate in machine units. If the commanded position is not reached the machine keeps moving even the jogwheel is not turning.
  
  Lead: Manipulates the spindle override.
  
  MPG: Manipulates the feedrate override.
  
  See here  xhc-whb04b-6  
  
    Signale, die zum Betätigen von Spindel und Vorschubübersteuerung verwendet werden. Der Vorschub-Drehknopf kann dienen als
  
• Countinou: Bewegung x% von der maximalen Geschwindigkeit
• Steps x mm
• MPG-Vorschubübersteuerung
• Die Sonderstellung Lead

         Continous: In diesem Modus wird das Joggen mit der gewählten Vorschubgeschwindigkeit durchgeführt. Wenn man                        das Jogwheel dreht, bewegt sich die ausgewählte Achse. 

          Step: In diesem Modus bewegt die Maschine mit den Schritten  wheel_counts bei der aktuell ausgewählten Schrittgröße                  und der aktuellen Soll Vorschubrate in Maschineneinheiten. Die Maschine bewegt sich so lange bis die Anzahl der Schritte                abgearbeitet ist, selbst dann wenn das Handrad nicht gedreht wird.
         Lead: Manipuliert die Spindelübersteuerung
         MPG: Manipuliert die Vorschubübersteuerung


              Siehe auch   xhc-whb04b-6

 

Exactly the same log message.
I also tested to unplug + plug back the dongle. Exactly the same log message.
Both ended up in the debug message in libusb.

It can of course be my hardware. It isn't to difficult to test with another test computer I have.

I use the built-in libusb-1.0 in debian bookworm (12.13)
It says 
#define LIBUSB_API_VERSION 0x01000109
in the include file.

I was able to install libusb 1.28, compile xhc component with it and at least it starts.
I have a minimal milling job this afternoon. I'll see how it reacts.

I made some changes in my firmware so I could generate a disconnect at will. Ten times in a row, I got libusb_submit_transfer to return non-zero in Usb::setupAsyncTransfer. And every time, the pendant reconnected to the xhc-whb04b-6 component without issue. In other words, I am unable to reproduce your issue, which more or less means I cannot help you directly. However, you mentioned that you had to change libusb_init_context() to libusb_init(). It appears libusb_init_context() was introduced in libusb 1.0.27:
libusb.sourceforge.io/api-1.0/group__lib...e12023b7399003c50470
Looking at my system, I have libusb 1.0.28-1. I think it would be well worth trying to upgrade libusb. Or if you don't want to do that, you could use LinuxCNC 2.9.8 as a "live cd" to see if that fixes the issue. This would actually be better, since it's possible that the issue could be in the Linux kernel itself.

What does the 'LEAD' function do?

Hallo Linuxcnc Freunde,

habe festgestellt das bei dem Handrad XHC-WHB04b-6 die Lead Funktion nicht richtig funktioniert.
Habe an der Treiberdatei pendant.cc die Zeile 346 hinzugefügt. Die "oder" Funktion für die 9b oder 1c wurde ja schon mit einegbaut. Danke an alle dafür.

Bei mir funktioniert diese jetzt und lässt sich aktivieren.
Hoffe konnte ein bisschen weiter helfen. Leider muss dazu Linuxcnc neu erstellt werden mit dem make Befehl.

Habe versucht die Datei mit hoch zu laden, hat aber nicht funktioniert.

[edit]
 

That was unexpected. For all the world, it looks like this get stuck in libusb somewhere. After the "is Open" debug text, the code does the following:
1. Function call XhcWhb04b6Component::enableReceiveAsyncTransfer()
2. From XhcWhb04b6Component::enableReceiveAsyncTransfer() call Usb::setupAsyncTransfer()
3. From Usb::setupAsyncTransfer() call libusb_submit_transfer()
4. From libusb_submit_transfer() call libusb_unref_device()
5. In libusb_unref_device() print "libusb: debug [libusb_unref_device] destroy device"
6. libusb_submit_transfer() never returns is my guess
Maybe some deadlock inside libusb. Whether that would be from some bug in libusb or because of some improper handling in the xhc-whb04b-6 component, I cannot say. At least at the moment. I'll look into it some more, but not right now.

Got a crash today. Seems to be the usual way. At mode and screen switch.I have seen that libusb debug printout several times before.Attach are the log and also usb.cc and xhc-whb04b6.cc to be able to follow my printouts. 

Okay. Once you get a crash, post the output here and we can try to work it through together. Please ensure that you have some print statement in place of the commented out assert call, so if it fails there, we know about it. Something like:

Still trying to learn exactly where in the code it fails.
It doesn't reconnect, that's for sure. If it had, it wouldn't have been an issue.
I overwrite the log and, alas, no saved useful log.
"Unfortunately" I didn't get a crash today when I worked with the machine.
I like the whb04b a lot.It's worth spending some time improving stability.

 

I was just in the garage using the mill with the pendant. And thought that maybe I should post a mood picture, so we don't forget why we put effort on improving the pendant software
 

Hakan, as far as I can tell, the component should try to reconnect after a failed asyncSetupTransfer in case you comment out the assert call. It should then after a short while fail on libusb_control_transfer, which then calls setDoReconnect(true), which in turn should cause a restart of the USB stack.

If you run the component with the -a flag, the console output when it fails would be interesting to see.