david@wucs2.UUCP (02/19/87)
Has anyone successfully used the SLIPDISC protocol (mentioned on page SMM:13-9 under tty_conf.c) to connect systems over an asynchronous serial link? We have the use of one channel of a u-wave link that appears as a 9600 baud serial line with occasional severe burst errors. We want to use the link to connect a uVaxII to a Vax-750 both running 4.3. We previously had used itt (posted by elvy@harvard in Feb 84) to (somewhat) succesfully connect a pair of 750's running 4.2 over the u-wave link. For reasons that we could never totally pin down, one of the machines crashed on the order of once a day when running itt. Converting itt from 4.2 to 4.3 does not seem to be straightforward. No one here has been able to succesfully compile a kernel with the SLIPDISC protocol. Does anyone know if it is functional as distributed, or if not know what modifications are needed. Any help or advice would be greatly appreciated. David Burstyn UUCP: david@wuibc2.UUCP Computer Systems Lab CSNET: wuibc2!david@seismo.ARPA%csnet-relay Washington University ARPANET: wuibc2!david@seismo.ARPA St. Louis, MO 63130 516 Lopata Hall, (314) 889-4678
chris@mimsy.UUCP (02/20/87)
In article <596@wucs2.UUCP> david@wucs2.UUCP (David Burstyn) writes: > Has anyone successfully used the SLIPDISC protocol >(mentioned on page SMM:13-9 under tty_conf.c) to connect >systems over an asynchronous serial link? Yes. (But not here.) > No one here has been able to succesfully compile a >kernel with the SLIPDISC protocol. Does anyone know >if it is functional as distributed, or if not know what >modifications are needed. It will run. Add pseudo-device sl 2 # or however many you need to your configuration file, run /etc/config, make depend, make. Boot the new vmunix, make sure characters sent on ttyXX show up at the other machine, and run slattach ttyXX 1200 & # or whatever speed If you need a modem dialled, write a program to do that, then run slattach. If the line drops, slattach should die with a SIGHUP, and you can automatically redial. There are, however, a few bugs in the code as distributed. In particular, if you get low on clists, the SL interfaces may lock up. I have no official fix yet (low priority). -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690) UUCP: seismo!mimsy!chris ARPA/CSNet: chris@mimsy.umd.edu