Closed parthy closed 11 years ago
I guess you're right. It's better to not have this detection and see (or not see at the beginning) pagefault-loops than having a wrong detection. Because in the end it does of course not work, since one could simply revoke memory and thus cause a fault on the same CPU and same address twice in a row.
The ChildManager detects that a pagefault was caused twice and kills the VM. But when I just disable this detection, everything works fine (i.e., no endless pagefault loop). I think I'd rather live with endless loops in the event of failure than having false aborts...