Clist problem

[Warren D. Swan]

In article <4922 at ncrcae.Columbia.NCR.COM> wescott at ncrcae.Columbia.NCR.COM (Mike Wescott) writes:
>In article <1026 at virtech.UUCP> cpcahil at virtech.UUCP (Conor P. Cahill) writes:
>> In article <134 at jassys.UUCP>, tony at jassys.UUCP (Tony Holden) writes:
>> > Question.  What exactly is clist (I know unix but not that deep).  CCI
>> > wants us to reboot every 3 days until they fix the problem.  Bogus?
>
>Actually there is a nasty little bug in SysVr2 clist manipulation code
>(fixed in SysVr3) in canon() in io/tty.c by which clist blocks can be lost.
>It is fairly rare since it requires two readers of the tty with ICANON not
>set.  The most common situation in which this occurs is running HDB uucp to
>dial out on a port that has uugetty running.

Aaaaaaaaaaaah.  A spark of insight.  Perhaps this might explain why the
system I use regularly (not cbnewsc) occasionally "hangs" (either everyone,
or just me, I forget).  It's running 5.2.6b (System V release 2 ...).
I and the aforementioned bug may be the culprits.

I use shl - the shell layer program for terminals that don't have layer
capabilities.  Since this program does the multiplexing of the input to
several shells (outputs), it seems like the subshells would be multiple
"readers of the tty with ICANON not set."

Furthermore, I've noticed for some time that it only hung when I was
using shl, or at least 99% of the time I was using shl (does anyone
else use it?).  'Cause I'd have to log back in and kill my processes
(including shl itself) left around when it hung.   Perhaps fodder for
comp.sys5.bugs?  Except we need to be sure before we publish such things
as fact there yet.

Thanks for your explanation.
Warren D. Swan  (WooDS)     Y n n ____ __      You can't tell which way a train
AT&T Bell Laboratories     -(((((([__]/__]     went by looking at the tracks.
Naperville, Illinois       /o-OOOOO~~  oo
att!cblph!woods         #####################  FRISCO 1630 Decapod (2-10-0) IRM