Date: Thu, 16 Nov 1995 14:42:20 +1100 From: Bruce Evans <bde@zeta.org.au> To: stesin@elvisti.kiev.ua, tateoka@pdcd.ilab.toshiba.co.jp Cc: hackers@freebsd.org Subject: Re: Is TTYHOG too small? (Re: STARTECH ...550CP chips are Ok for me. ) Message-ID: <199511160342.OAA32756@godzilla.zeta.org.au>
next in thread | raw e-mail | index | archive | help
># I found a solution to my problem with UUCP i protocol. ># It's not in the sio.c but tty.h. ># ># after 2.0.5, clist size limited by c_cbmax. ># and c_cbmax's value from TTYHOG in tty.h. ># but flow control refer to t_hiwat from TTYMAXHIWAT. ># and TTYHOG < TTYMAXHIWAT. ># ># I think It cause unusual flow control ( by no more clist ). TTYMAXHIWAT is essentially the size of the output buffer. It is unrelated to TTYHOG. TTYHIWAT is currently 2048 while TTYHOG is only 1024. These numbers haven't changed for a long time and made more sense when most i/o was to terminals so there was always a lot more output than input. ># Last night, I increased TTYHOG to 4096 in 2.1.0-951104-SNAP and ># recompiled my kernel. ># I tryed UUCP communication, and uucico completed i protocol transfer. Making TTYHOG larger may help solve flow control problems. Hardware flow control should always work, but throught may be reduced a little if flow control is invoked. Are you using hardware flow control in both directions? I think uucp's `hardflow' option only applies to outgoing connections. To initialize crtscts flow control for uucp logins (and all other logins), use `stty crtscts' on the initial state ports and perhaps on the lock state ports for each dialin port. See /etc/rc.serial. Bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199511160342.OAA32756>