We've seen reports of VisualWorks crashing on 2.6.x (x86) kernels with a segmentation fault. I've just gotten a message from engineering that pinpoints the problem:

There's a kernel patch floating around that implements a kernel barrier to executing code at addresses that fall into the stack or data segments of the process. The intention is to provide a safety net for stack frame overflow attacks. The 2.6.8-1.521 kernel from FC2 uses this patch, as well as several others that muck around with the exec shield system. Since VW runs on FC2 using 2.6.7-1.494.2.2 which also contains the exec shield patch, I suspect that one of the suplementary patches is causing this problem.

This exec check can be globally disabled by executing the following as root:

#> echo 0 > /proc/sys/kernel/exec-shield

Now, we've also heard that the JVM has issues on the 2.6.x kernels; it's likely the same kind of problem. We'll have to address this in the future, but for now, try the suggestion above.