Is the threaded one coglinuxht?  I'm 99% sure I'm using that because I
want to take advantage of any extra processing-cores if I can.  Here's
my -version output just FYI:

4.0-2776 #1 Thu Aug 22 10:36:42 PDT 2013 gcc 4.1.2 [Production VM]
CoInterpreter VMMaker.oscog-eem.331 uuid:
37d2e4b0-2f37-4e2d-8313-c63637785e59 Aug 22 2013
StackToRegisterMappingCogit VMMaker.oscog-eem.333 uuid:
84da9cb8-7f30-4cb7-b4fb-239a11f63b54 Aug 22 2013
VM: r2776 http://www.squeakvm.org/svn/squeak/branches/Cog
Plugins: r2545 http://squeakvm.org/svn/squeak/trunk/platforms/Cross/plugins
Linux mcqfes 2.6.18-128.el5 #1 SMP Wed Jan 21 10:44:23 EST 2009 i686
i686 i386 GNU/Linux

>> I try with the -memory parameter instead of -mmap, then Cog crashes
>> immediately with -memory 2200m when launching the stock 4.5 image..
>
> Yes, that's your limit with X11.  32-bit linux reserves the top 1Gb for
> kernel space, leaving 3Gb for user processes.  Once you factor in space
> taken for the VM and the X11 libraries (plus sundries) 2200mb is simply too
> much memory and the system crashes in X11 startup.  So for the moment you'll
> need to peg things at 2100mb or 2000mb.

Ok, but what about the reproducible case I documented?

>> > You can easily reproduce it yourself with Cog.  Just start the base
>> > Squeak4.5 image with the -mmap 2000m argument.
>> >
>> > Then, in a workspace:
>> >
>> >    self halt.
>> >
>> > Do It, then press Proceed on the debugger that appears.
>> > Does the problem occur for you too?  It's related to amount reclaimed
>> > by the garbage-collector, but clearly we're not out of memory....

Is this a problem?

Thanks.