[net.unix-wizards] Reboot problems on 750

jpd@its63b.ed.ac.uk (P Dourish) (10/17/86)

On EUSIP, a VAX-11/750 running 4.2, we have a strange boot problem. It doesn't
do the machine any harm (I don't think it does, anyway) but it's annoying.
Whenever the VAX boots, it hangs between the autoconfiguration and the start
of the rc file. It will refuse to begin the rc file until someone types
Return on the console; once that happens, it goes happily on its way.
Here's a boot log:

>    %%
>    loading hp(0,0)boot
>    Boot
>    : hp(0,0)vmunix
>    170624+49728+40804 start 0xfa4
>    4.2 BSD UNIX #4: Wed Jul 9 10:25:40 WET 1986
>    real mem  = 8384512
>    avail mem = 7085056
>    using 183 buffers containing 838656 bytes of memory
>    mcr0 at tr0
>    mba0 at tr4
>    hp0 at mba0 drive 0
>    hp0: eagle
>    hp1 at mba0 drive 1
>    hp1: eagle
>    hp2 at mba0 drive 2
>    hp2: eagle
>    uba0 at tr8
>    zs0 at uba0 csr 172520 vec 224, ipl 15
>    ts0 at zs0 slave 0
>    dh0 at uba0 csr 160020 vec 300, ipl 15
>    dm0 at uba0 csr 170500 vec 310, ipl 14
>    de0 at uba0 csr 174510 vec 120, ipl 15
     ---===<<<((( HANGS HERE UNTIL "RETURN" )))>>>===---
>    Automatic reboot in progress...

I'd be grateful for any suggestions anyone can make as to why it's happening
and how we might be able to fix it.

Thanks.

Paul Dourish, Speech Input Project, University of Edinburgh

                   ARPA:  jpd%ed.eusip@ucl-cs.arpa
                   UUCP:  ...!seismo!mcvax!ukc!eusip!jpd
                   JANET: jpd@uk.ac.ed.eusip

"The American fondness for Nuclear Weapons is very patriotic. After all, it
is enshrined in their Constitution that all shall be cremated equal..."

thomas@utah-gr.UUCP (Spencer W. Thomas) (10/19/86)

This question has already been answered several times in this group, to
my knowledge.  Apparently it bears repeating again.  The problem is that
the console tty is running slower than the output data is being sent to
it.  When its internal buffer fills up, it sends a ^S, later it sends a
^Q.  Since the tty driver hasn't been "started" yet, the ^S is still in
the (hardware) input buffer and the ^Q is dropped.  When the console
driver becomes active it reads the ^S and dutifully stops output until a
character is typed.

A solution is to lower the baud rate of your console interface or to get
a console that can print faster.

-- 
=Spencer   ({ihnp4,decvax}!utah-cs!thomas, thomas@utah-cs.ARPA)