[comp.sys.att] tty000 'n' ph0 'n' BREAK

ford@crash.CTS.COM (Michael Ditto) (10/04/87)

Further testing with a small C program to do the TCSBRK on ph0 revealed
that doing the TCSBRK on /dev/ph0 DOES cause tty000 to become wedged.  There
was no response to keys typed on tty000, including BREAK.  Output to tty000
(initiated by "echo foofoofoofoo > /dev/tty000" on the console) buffered up
for a few hundred characters or so, but then would block.  Stty could read
the termio settings of tty000 (via "stty < /dev/tty000") but attempts to
change them (via "stty -ixon < /dev/tty000") would hang indefinately.  What
finally made tty000 come back to life was "stty -ixon -ixoff < /dev/ph0".
After that, tty000 worked fine.  I send a TCSBRK to ph0 again, this time
from the console, and again, tty000 froze.  Again, "stty -ixon < /dev/ph0"
made tty000 come back.  Note that ph0 was ALREADY in -ixon mode, so it was
simply a TCSETA to ph0 with no changes that made tty000 comp back.  I did
notice that when it came back, pending output did come out, but the keys I
had typed on tty000 while it was frozen never showed up.

Note that I ran these tests with kernel 3.51.1, (the 'tty000 break'
fix) although these problems did happen with the normal 3.51 kernel.

As I narrow down the problem further, hopefully it will reach a point where
I have enough information to make a call to the "hotline" worthwhile.  Send
me any other details you discover...

-- 

Michael "Ford" Ditto				-=] Ford [=-
P.O. Box 1721					ford@crash.CTS.COM
Bonita, CA 92002				ford%oz@prep.mit.ai.edu