1. kernel: drivers/usb/input/hid-core.c: input irq status -84 received
That particular code indicates either that the Belkin adapter sent an illegal bit stream or that it failed to respond at all to a request.
Okay, that makes sense. Maybe it was the CueCat that sent the wrong stream or failed to respond correctly. More below...
2. The kernel message above comes whenever I scan a bar code with my CueCat bar code reader which is connected to a Thinkpad T42 through a Belkin USB to PS/2 Adapter (F5U119-E). Some characters from the scan are transferred to the xterm or gnome-terminal that has focus, but then the laptop keyboard locks up (and mouse behaves erratically). To get control back, I have to disconnect the Cuecat and Belkin USB to PS/2 adapter and wait. This happens even when stdin is redirected to a file. I may have an unusual combiniation of hardware. If there is anything I can do to help out (like running test patches), I would love to.
Do things work any better when you don't work under X?
Not much better. It took two or three swipes with the CueCat but then a stream of repeating characters began. I had to Ctrl-C. Aside from Ctrl-C, the keyboard was pretty much locked up until I disconnected the USB to PS/2 adapter. I think I had to even use C-A-D to reboot. C-A-D worked, and numlock toggled, but regular keys would just beep at me. The two or three swipes that I did partially worked in that they produced some output but it didn't look proper.
The Thinkpad T42 does not have PS/2 ports, so I am using the Belkin adapter. I have another computer also running Debian Testing which is just a regular computer, not a laptop, which has a PS/2 keyboard. On that computer, the CueCat works as expected. I think this problem is related to the USB driver not handling the combination of Belkin adapter and Cuecat. When I plug a regular keyboard into the Belkin adapter and connect that to the Thinkpad, the external keyboard works fine.
On the Thinkpad, when the Belkin adapter is connected, the following are logged to syslog:
kernel: input: USB HID v1.00 Keyboard [USBPS2] on usb-0000:00:1d.1-2 kernel: input: USB HID v1.00 Mouse [USBPS2] on usb-0000:00:1d.1-2
It seems a little strange that the one adapter appears as two devices, but I'm not thoroughly familiar with the USB HID layer.
I tried two other USB to PS/2 adapters today. One by Compaq looked just like the Belkin but rebranded. It had two PS/2 ports, color coded to keyboard and mouse with a single USB plug for connection to the computer. When I connected the Compaq adapter, it produced the same entries (USBPS2 Keyboard and Mouse) in syslog.
Radioshack makes a single PS/2 port to USB adapter, but when I connected that one, nothing at all was recorded in syslog and swiping the same bar code with the Cat produced no output.
Nothing additional is logged when the Cuecat is connected to the keyboard PS/2 connected on the Belkin adapter. (Aside from the warning/error message about IRQ status after scanning a barcode.)
I suspect the problem is an incompatibility between the adapter and the Cuecat; probably the adapter expects either a keyboard or a mouse to be plugged in but the Cuecat doesn't quite behave like either one.
That may be it. I think the Cuecat normally behaves like a keyboard. It connects inline with a PS/2 keyboard. It has M and F PS/2 connectors, so you can plug a keyboard in to the CueCat F PS/2 connector and then plug them both into the computer's PS/2 port. The Belkin packaging says "Supports nearly all popular PS/2 input devices." After more testing, the Belkin may live up to its statement...
* * *
My comments from the first e-mail, and comments here with regard to things not working all occured while connecting a CueCat with Cat. No. 68-1965-A though the Belkin or Compaq USB to PS/2 adapter. That CueCat worked when plugged directly into a PS/2 port, but not through an adapter.
I dug up another CueCat, this one has Cat. No. 68-1965 (no -A) and it works fine though the Belkin adapter. I don't know if it is newer or older than the -A model, but is definately behaves differently though the USB adapter.
What happens if you use the adapter/Cuecat combination with your regular computer's USB port?
With the original CueCat that I was using, it's even worse than on laptop. When I connected it, everything looked the same in syslog, but as soon as I swiped something, the computer got into a loop repeating the bell from the PC speaker. I hit Scroll Lock to see if I could toggle the light on the keyboard, this cause something that looked like memory information to dump to the screen (it was not an OOPs though.) I tapped Scroll Lock a few more times and then the keyboard completely locked and I lost video too. Weird. I had to power cycle to reboot.
In testing, I also tried the CueCat (the -A one) though Windows XP on my Thinkpad through the Belkin adapter. With Windows XP the results were almost the same. Swipes of bar codes would produce some output but not the proper output. After that, the keyboard (in Word Pad and the command shell) was "funny". Caps lock was inverted. The main difference between Windows and XP was that Windows didn't lockup after swiping a bar code.
Hopefully I closed the loop for anyone interested in this. I'm happy because I found a working CueCat. It appears the different model numbers differ in their ability to work with Belkin and Compaq USB to PS/2 adapters.
I'm suprised that the -A model is able to crash the kernel and cause a lockup. I guess if that CueCat is acting faulty, that's possible.
If anyone is interested in patching the USB drivers involved and wants me to run anymore tests, I'd be happy to.
Thanks again! Brian
------------------------------------------------------- The SF.Net email is sponsored by: Beat the post-holiday blues Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek. It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt _______________________________________________ [email protected] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
