On Thu, 26 Aug 2010, Bill Pemberton wrote: > I've run into a problem with usb-serial that has me puzzled. I've got > one particular machine where usb-serial devices just don't work. The > machine does correctly deal with other USB 1.1 devices, such as a > keyboard, so it's not just USB in general not working. The machine > uses a MSI 760GM-E51 motherboard with the AMD SB700. The problem is > that running the serial_loopback perl script will fail at the call to > getlflag(). > > I get the same failure using 3 different usb-serial converters: 1 is > pl2303, 1 is keyspan, and 1 is Quatech SSU-100. Also I've tried it > with several different kernel versions including 2.6.35 and > 2.6.36-rc2. These combos work on other machines I've got, it's just > the SB700 based machine where it fails. > > I'll attach 2 usbmon output files in case some one can spot a problem > using them. The difference between the good and the bad seems minor > to me. These are both using a pl2303 based adapter. The .good > output was on a AMD-8111 based machine the .bad output is from the MSI > machine. In each case the usbmon was run with no other USB devices > plugged into the machine. The pl2303 was plugged in, > serial_loopback.pl -e was run, the pl2303 was unplugged, usbmon was > stopped. The main difference between the two files is that the .good run sets the baud rate to 9600 and the .bad run sets it to 115200. There's also a difference later on; the .good file shows three '!' characters being sent to and read back from the device whereas the .bad file shows no corresponding I/O. I don't know the reason for the difference in baud rates. Did both machines have the same version of the kernel and the same Perl program file? 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
