**** BEGIN LOGGING AT Sun May 03 02:59:58 2009
May 03 12:34:17 <Strontium> anybody here know anything about the flyswatter?
May 03 14:58:57 <keesj> Strontium: what is your question?
May 03 15:04:59 <keesj> Strontium: what is your question?
May 03 15:06:39 <Strontium> I wondered if it is normal to have problems using both OpenOCD and the on board serial port of the flyswatter (uisng /dev/ttyUSB1) at the same time.
May 03 15:10:15 <keesj> "normal" is a little strange in this context. It is know that if you first connect to the board via serial and only after that start openocd the serial goes boom
May 03 15:10:35 <keesj> but you can still "reconnect" afterwards , are you have more the the "normal" problem?
May 03 15:10:55 <Strontium> yes, i cant reconnect after
May 03 15:14:50 <keesj> Somebody (you?) mailed about this yesterday on the mailing list
May 03 15:15:12 <Strontium> yep
May 03 15:15:27 <keesj> I don't think this is normal behaviour(at least I think it should be possible to modify openocd to not do this)
May 03 15:16:11 <Strontium> i have since hacked up a quick and dirty program to dump serial data from the port, using libftdi, but it doesnt work for port B, only port A.  I get errors in dmesg if I try port B
May 03 15:16:36 <Strontium> I think the error is something to do with libftdi and libusb interaction
May 03 15:17:01 <Strontium> I think when you open the ft2232 with libftdi it just grabs the whole device. (both ports)
May 03 15:17:23 <roxfan> have you tried the other ftd lib?
May 03 15:18:06 <Strontium> well yes, and no.  I didnt get much sense out of it, so didnt persevere.  I should go back to it and do more experiments.
May 03 15:18:35 <keesj> booth have the problem 
May 03 15:18:54 <keesj> I don't know if libusb even allows this
May 03 15:20:00 <Strontium> I will keep experimenting.  I am currently about to install the latest libftdi from source (rather than the one that ubuntu ships)
May 03 15:20:20 <Strontium> That way I can put some tests into the library to try and work out what is going on
May 03 15:23:41 <keesj> probably this:
May 03 15:23:42 <keesj> src/ftdi.c:    if (usb_detach_kernel_driver_np(ftdi->usb_dev, ftdi->interface) != 0 && errno != ENODATA)
May 03 15:27:09 <Strontium> A quick check would indicate if interface is correct, then it should only detach the driver for ttyUSB0
May 03 15:27:14 <keesj> (this is libftdi-0.15)
May 03 15:29:38 <Strontium> although looking at that code showed me my bug.  I was calling set interface AFTER the open.  It needs to be called before.
May 03 15:31:13 <Strontium> OpenOCD does the same thing
May 03 15:32:57 <Strontium> Although in OpenOCD's case, it should be not a problem, because interface A is the default
May 03 15:33:02 <Strontium> still
May 03 15:43:01 <Strontium> I will link up my test program which can simulate the openocd problem and trace it through GDB to find out which particular usb operation kills Interface B tomorrow.
**** ENDING LOGGING AT Mon May 04 02:59:57 2009