larry@mitra.mitra.com (Larry Williamson) (04/19/91)
It seems to be quite simple to panic Ultrix 4.1 on a DEC station 5000 using dbx. Dbx a process that does an accept(). While your process is blocked on the accept() call, type ^C to terminate the process. Dbx responds telling you your process has terminated, now quit dbx. Voila, panic. This may be something peculiar to our system, but over the course of the last few days we have done this 5 or 6 times with consistant results. -Larry
fortin@zap.uucp (Denis Fortin) (04/21/91)
In <LARRY.91Apr19073000@mitra.mitra.com> larry@mitra.mitra.com (Larry Williamson) writes: >It seems to be quite simple to panic Ultrix 4.1 on a DEC station 5000 >using dbx. >Dbx a process that does an accept(). While your process is blocked on >the accept() call, type ^C to terminate the process. Dbx responds >telling you your process has terminated, now quit dbx. Voila, panic. >This may be something peculiar to our system, but over the course of >the last few days we have done this 5 or 6 times with consistant >results. Arrrrrgh! This is *not* peculiar to your system. We have been experiencing the same problem on our DECsystem 3100s running Ultrix 4.0! (and 3.x before that). It has gotten to the point that we tell our developpers to "use printfs instead of using dbx" (!). Sigh... Any workaround for this problem would be greatly appreciated. -- Denis Fortin, DMR Group Inc, (514) 877-3301 (All of these opinions are my own) fortin@zap.uucp uunet!sobeco!zap!fortin fortin%zap@larry.mcrcim.mcgill.edu
mjr@hussar.dco.dec.com (Marcus J. Ranum) (04/22/91)
>Dbx a process that does an accept(). While your process is blocked on >the accept() call, type ^C to terminate the process. Dbx responds >telling you your process has terminated, now quit dbx. Voila, panic. I just tried this on a 3100, running: ULTRIX V4.1 (Rev. 52) System #2: Thu Apr 11 16:23:19 EDT 1991 and it survived just fine. If someone can show me how to reliably reproduce the problem, I'll QAR it, but I can't until it's something I can duplicate. mjr.
neideck@cloris.enet.dec.com (Burkhard Neidecker-Lutz) (04/22/91)
It was a strange kernel bug that is fixed in Ultix 4.2. Bit us too... Burkhard Neidecker-Lutz, CEC Karlsruhe