Return to BSD News archive
Path: euryale.cc.adfa.oz.au!olive.mil.adfa.oz.au!navmat.navy.gov.au!posgate.acis.com.au!warrane.connect.com.au!news.syd.connect.com.au!gidora.kralizec.net.au!not-for-mail From: bde@zeta.org.au (Bruce Evans) Newsgroups: comp.unix.bsd.freebsd.misc,comp.mail.uucp Subject: Re: Taylor-UUCP not running with 16550 SIO (UMC Custom Chip) Date: 26 Feb 1996 20:36:06 +1100 Organization: Kralizec Dialup Unix Lines: 34 Message-ID: <4gruu6$uhq@godzilla.zeta.org.au> References: <DMonLF.Ctq@citylink.dinoex.sub.org> <DMquCt.H65@nemesis.lonestar.org> <DnD0Lz.CFw@citylink.dinoex.sub.org> NNTP-Posting-Host: godzilla.zeta.org.au Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:14395 comp.mail.uucp:15849 In article <DnD0Lz.CFw@citylink.dinoex.sub.org>, Peter Much <peter@citylink.dinoex.sub.org> wrote: >In article <DMquCt.H65@nemesis.lonestar.org>, >Frank Durda IV <uhclem@nemesis.lonestar.org> wrote: > >>[0]bps. This is baudrate-independent, it happens with 300 baud as with 57600 >> >>Here is the patch I apply every time I upgrade FreeBSD. It fixes >>Taylor UUCP to actually use hardware flow control for incoming and >>outgoing calls. Without it, I get terrible UUCP results. With the fix > >Thanks, but hardware-handshake wasn't the source of the described trouble, >otherwise it would have worked with 300/1200/2400 Baud - these don't need >handshaking. >I think it's timing problems in the UMC Chip's SIO. Yes, mine loses sync if the fifo enable bit is toggled or the speed is set while part of a character has arrived. The UART apparently syncs with the wrong start and stop bits and doesn't recover until data stops arriving or the UART is reprogrammed while it isn't in the middle of receiving a character. This happens at all speeds that I tried (between 300 and 115200) under all OS's that I tried (FreeBSD and Linux). I have some partial fixes. E.g., never changing the fifo enable bit, and only setting the speed if the speed has changed should work in most cases. Currently the speed is always set when the termios settings are set (even if it hasn't changed), and the fifo enable bit is only changed to fix the a fifo bug in an SMC custom chip and to fix bugs in BIOS probes. The problem is probably more visible for uucp than for ppp and zmodem because uucp changes the termios setting more. -- Bruce Evans bde@zeta.org.au