On Fri, 24 Sep 2010, Søren Holm wrote: > Onsdag 22 september 2010 22:50:33 skrev Alan Stern: > > You should modify the pl2303.c file by adding a "#define DEBUG" line > > before any of the "#include"s. Then see what shows up in the dmesg > > log. Maybe also collect a usbmon trace at the same time. > > Ok, I tried compiling with debug defined above the includes. Nothing other than > the usual ended up in the logs. Did you check the dmesg log? Did you make sure that the modified driver module was loaded? > I tried logging data using usbmon. I did two runs both containing the trafic > when doing the following. > > 1. Plug in of the device > 2. Execute "cat /dev/ttyUSB0" on the device > 3. Stop the cat-process using CTRL-C > > I first created pl2303_not_working.log from a fresh booted system. This is the > scenario when the open-call is blocking. > > Then I unplugged the device and issued these commands : > # rmmod pl2303 > # rmmod usbserial > > Then I produced pl2303_working.log which is the scenario when the open-call is > not blocking. > > I hope it help trackig down the bug. The logs do show a difference. It involves the baud rate; the not_working log sets a baud rate of 115200 whereas the working log sets a baud rate of 9600. That's the problem with using "cat" on a serial port -- you have no control over the port settings. Instead you should use a program designed for serial ports, like minicom. I don't know why the higher baud rate caused the open call to block. And I don't know what caused the difference in baud rates. It could be some other program completely, like a network manager trying to use the serial port for a network link. > Btw. Shouldn't the communication be done via. bugzilla ? If you want. You started this conversation by posting to linux-usb. Instead you could have added linux-usb to the bug report's CC list and held the conversation that way. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
