From: Jonathan Bredin <jbredin@co...> - 2005-04-06 17:21:54
Thanks for the pointers. Your SerialPort.cpp code does indeed work, and I have a C version as well now. I ran an strace of the Acroname code and my code to find the only difference is that my code calls an extra fcntrl... though I cannot see where it comes from comparing the two sources.
/dev/ttyS0 is the port I want.... as long as I'm not logged into it myself ;)
Dave Hylands wrote:
> Hi Jonathan,
> > I've been logging in over bluetooth to a gumstix connected with a serial
> > cable (not a null modem) on the leftmost (ttyS0) socket connected to the
> > Brainstem.
> How are you initializing the serial port on the gumstix side of things?
> I have some code that *should* work (it works fine under cygwin).
> Checkout: http://www.davehylands.com/Software/BootHost/SerialPort.cp
> This page has lots of good info:
> Make sure that you choose the right mode (canonical - or cooked mode
> is typically used for processing line based input and control
> characters have special meaning, non-canonical or raw mode is for
> binary style input and a control character is just 8 more bits of
> Remember that ttyS0 is normally the console port (where all the output
> goes from the kernel during bootup).
> Dave Hylands
> Vancouver, BC, Canada
From: Jonathan Bredin <jbredin@co...> - 2005-04-06 17:31:29
Thanks, Craig. I was concerned about interfering with Bluetooth/bnep
(on /dev/ttyS1?). I'll try switching the bnep.... hmm I see it's
already on ttyS3. Good.
Craig Hughes wrote:
> On Apr 5, 2005, at 3:14 PM, Dave Hylands wrote:
> > Remember that ttyS0 is normally the console port (where all the output
> > goes from the kernel during bootup).
> Not just bootup -- even after all the boot messages, if you"re booting
> with console=ttyS0 then any other console messages after bootup (like
> application error messages and such) might also go to that device. Any
> reason you wouldn"t use the other serial port in this case? Or are you
> already doing that for some other device?