Greg, Thank you for your reply. It's also reassuring to know a bug was actually found and *fixed*! 8-)
I have some news. I tried a Belkin F5U109 and it worked, at least for the one device I was trying to get working. I did not experience a dropped character after 20 trials. Here what I'm seeing in /var/log/messages for your information: Sep 23 12:55:24 Grisham kernel: hub.c: USB new device connect on bus2/2, assigned device number 5 Sep 23 12:55:24 Grisham kernel: usb.c: USB device 5 (vend/prod 0x50d/0x109) is not claimed by any active driver. Sep 23 12:55:27 Grisham /etc/hotplug/usb.agent: Setup mct_u232 for USB product 50d/109/102 Sep 23 12:55:27 Grisham kernel: usbserial.c: USB Serial support registered for Magic Control Technology USB-RS232 Sep 23 12:55:27 Grisham kernel: usbserial.c: Magic Control Technology USB-RS232 converter detected Sep 23 12:55:27 Grisham kernel: usbserial.c: Magic Control Technology USB-RS232 converter now attached to ttyUSB0 (or usb/tts/0 for devfs) Sep 23 12:55:27 Grisham kernel: usbserial.c: USB Serial support registered for MCT/Sitecom USB-RS232 Sep 23 12:55:27 Grisham kernel: usbserial.c: USB Serial support registered for MCT/D-Link DU-H3SP USB BAY Sep 23 12:55:27 Grisham kernel: mct_u232.c: v1.1:Magic Control Technology USB-RS232 converter driver [Now that I look for it in the "working devices list," I see it is mentioned there as working with kernel 2.4.18 as well. That's what I get for searching for "serial" and not "adapter".] I have a follow-up I hope you can comment upon. Now that it seems the first device works and doesn't drop characters, I have one other device that I'd like to see if I can get working (but it's not crucial) with the same adapter. I'm forced to use the adapter as the target laptop only has one serial port. The second device works fine with /dev/ttyS0. Should the code that works with the serial port be able to work unchanged with /dev/ttyUSB0? My guess is that it should. The current code sets line speed (B19200) and mode (CS7 | CLOCAL | CREAD | PARENB), all that serial stuff you would expect; may I presume the com- bination of serial-to-USB adapter and driver should cause the adapter to be re- configured to use the correct serial settings? If all of this is true then it's curious that this just doesn't work. Perhaps I've found a problem. I would be glad to pursue it if it would ultimately help to shore up a bug. If you can give me any pointers on proceeding on debugging this situation I would be glad to try. It might be useful to know the devide in question only trans- mits data. Thanks again for the feedback on the previous question. If the Belkin F5U109 hadn't work we'd likely be forced into a kernel upgrade. However, given the issues with 2.4.18-7 to surely appears we should ugrade the kernel anyway! Regards, Joel Breazeale > On Tue, Sep 23, 2003 at 12:40:16PM -0500, Joel L. Breazeale wrote: > > > > Can anyone out there give me a recommendation on a manufacturer that has a > > > > serial-to-USB adapter that works? > > > > > > The Edgeport and Keyspan devices _do_ work, if you don't insist on using > > > a known buggy and insecure kernel that is quite a few years old... Oh > > > yeah, there have been a _lot_ of USB fixes since that kernel release > > > too. > > > > I'll pass along your comment. I believe the latest kernel for RH 7.3 > > is 2.4.20-20.7. I should think that would be sufficient for the > > Edgeport and Keyspan to work without kernel modification/recompile, > > yes? > > Yes, they should work just fine. If not, please let me (and Red Hat) > know. > > > > > [I've checked the LiNUX-USB device list. That's how I found the above devices. > > > > I'd really like to get someone with personal experience to respond. In the > > > > mean- > > > > time I'll keep checking items in the device list.] > > > > > > How about a dirt cheap pl2303 based device. That might work for your > > > application (although I would recommend a new kernel even for that > > > device...) > > > > For some reason PL2303 (no matter what manufacturer) has the chance of > > dropping the first character on the first read. My gut says is between > > one in five and one in ten the first character will be dropped, so using > > a PL2303-based device seems to not make sense. > > Yeah, I _just_ fixed this bug in the driver. It will show up in the > 2.4.23 kernel release. > > > At this point we're going to have to seriously consider using a more > > recent kernel. > > I think that is your best solution. > > Good luck, > > greg k-h > ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ [EMAIL PROTECTED] To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users
