Date: Sun, 2 Mar 1997 09:27:49 -0800 (PST) From: "Eric J. Schwertfeger" <ejs@bfd.com> To: freebsd-hackers@FreeBSD.ORG Subject: sio projects Message-ID: <Pine.BSF.3.95.970302091248.18611E-100000@harlie.bfd.com>
next in thread | raw e-mail | index | archive | help
Before I embark on my first FreeBSD kernel hacking project, does anyone know if someone is already working on 16650/16750 UART support? My plans are for the following 1) Autodetect the UARTS, and use their deeper fifos. 2) Allow for the software-selectable divisor bit, for baud rates above 115k. 3) Throw in a flag to use the UART-driven hardware flow control methods. 4) Throw in a flag to tell the driver that the clock crystal has been replaced (for those trying for more than 115K with 16550s). Nothing important, but the system really should know the real baud rate, if for no other reason that the user doesn't have to remember to adjust it in his head. Also, I've found the info for #2 for the 16650, and some 16650 serial cards (should arrive this week), but I would appreciate any pointers to full data sheets on both UARTS, and a source for the 16750 UART chip. Finally, my next project after this is to work on kernel PPP. For this project, I'm setting up two 486/33s side by side, connected by both a null modem cable and ethernet, with one sysloging to the other, so I should have a very good idea where things are going wrong. Currently, with kernel PPP, I'm seeing two problems. The first is that every once in a while, while the modems haven't dropped carrier, packets stop making it across teh PPP link. Second, at least with Linux PPP, lcp-echo-failure doesn't work, as the remote end drops the FreeBSD end due to not seeing any lcp-echos. Not a problem between two linux boxes, so I suspect the problem is in FreeBSD.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.95.970302091248.18611E-100000>