On Dec 17,  7:10am, Calvin wrote:
> On Sat, 16 Dec 1995, Jake Hamby wrote:
...
> > This test program which will reliably segfault with i=174687, which
> > you can verify with gdb on the core.  It creates 500,000 nodes in a chain,
> > then traverses the chain, then recursively traverses the chain and frees
> > the nodes in reverse order (it is when traversing the second time,
> > searching for the end node, that it segfaults).  Considering the large
> > number of nodes, it should give a good workout to the VM subsystem as well
> > as malloc/free/sbrk.  It consistently crashes on my NetBSD/Amiga system.
...
> havn't tested it on my system but how about this... if you're running CSH:
> 
> limit stackspace 65535
> 
> this will increase your stack from 8K to 64K.  This should solve your 
> problem.  If this doesn't work, try:
> 
> unlimit stackspace
> 
> and run your program again.  The default stackspace csh gives to a 
> program is 8K and with an application like yours, it seems like you're 
> using up all your stack quite quickly.  When a program runs out of 

  I did test this :-)

  I'm running bash, and the default user stack was the same as the NetBSD
defined default stack, which is 2MB.  I increased this to 4MB and still got
a segement violation, but the program run sucessfully with the user stack
size at 8MB.

Michael

-- 
Michael L. Hitch			INTERNET:  osymh@montana.edu
Computer Consultant
Information Technology Center
Montana State University	Bozeman, MT	USA