- From: bml <bml22_at_cornell.edu>
- Date: Sat, 03 Aug 2002 01:53:11 -0400
I'm having trouble properly terminating a command from both kernel space and user space. I'm using the ni_pcidio driver with the latest stable version of kcomedilib and comedi. I've modified the driver to respond to an external timer and use the stop_src TRIG_NONE. Everything works fine with the data collection, but when I try to stop the command using command_cancel(), the response is strange. In kernel space interrupts keep occuring, although my callbacks don't occur anymore, however the callback event mask seems to be unchanged. In user space i'm simply seeing the interrupts continue to occur. As strange as it seems, in user-space everything *appears* to be functioning fine: when comedi_cancel() is called, read() will return -1, but the interrupts don't stop. A near term fix has been running demo/cmd.c (for example) with stop_src = TRIG_COUNT instead. This stops the interrupts. Is TRIG_NONE as a stop_src working on any driver? Ideas? ---brent
Received on 2002-08-03Z04:53:11