On Monday, March 18th, 2024 at 12:36 PM, Dave Horsfall <dave(a)horsfall.org> wrote: ttys(V) in the Sixth Edition indicates the first digit of an /etc/ttys entry indicates a terminal line is active on init and the second indicates the final character of the /dev/ttyx entry[1]. Looking at both the Fifth and Sixth Edition /etc/ttys in the archive[2][3], both only have a 1 in the first column of entry 8, corresponding with /dev/tty8. From the setup document distributed with the Sixth Edition[4]: "The same goes for the character devices. Here the names are arbitrary except that devices meant to be used for teletype access should be named /dev/ttyX, where X is any character. The files tty8 (console), mem, kmem, null are already correctly configured." From all of this it appears that by convention, tty8 was indeed the default console /dev entry, although this could be changed by editing the conf and ttys entries followed by regeneration of the system. This would change in the Seventh Edition with the rearrangement of the ttys file to indicate longer /dev entry names and the establishment of a specific /dev/console entry. - Matt G. [1] - https://www.tuhs.org/cgi-bin/utree.pl?file=V6/usr/man/man5/ttys.5 [2] - https://www.tuhs.org/cgi-bin/utree.pl?file=V5/etc/ttys [3] - https://www.tuhs.org/cgi-bin/utree.pl?file=V6/etc/ttys [4] - https://www.tuhs.org/cgi-bin/utree.pl?file=V6/usr/doc/start/start

On Mon, Mar 18, 2024 at 9:57 AM Clem Cole <clemc(a)ccc.com> wrote: I'm pretty sure that when we beta tested the VAX-11/780 in the summer of 1977 it had a DZ11. We only had a few LA36 printing terminals on it so we never saw any issues with it. Not so when I started work at DEC in 1980. The VAX-11/780 I worked on had IIRC three DZ11a supporting 24 terminals. The lack of DMA was a big issue. We had to run our VT52s no faster than 300 baud or it brought the system to its knees. -Paul W.

The other problem with the DZ is it was one interrupt per character if I recall. Thte DH you could get multiple (output) characters per interrupt. Greatly decreased the load on the system. I do remember the KL kernel prints were not interrupt driven so the system pretty much froze while the kernel printfs were being output. There was a comment on the code saying this was “Not for idle chit chat.”

On Mar 18, 2024, at 1:21 PM, segaloco via TUHS <tuhs(a)tuhs.org> wrote: NatSemi (later Ti) 16550s have a 16 byte fifo in each direction. Any delays would be in getc(), putc(), used for system console io, not general serial io. Ignoring any code for brokenness! The early 16660 versions had a bug preventing use of the fifo but this was fixed ages ago. At least this is what I remember decades later!