the performance impact applies to kernels running in a virtual machine only, on native hardware there should be nothing measurable.ralphy wrote:hello everybody i'm curious as to whether anybody has tested the new pax test patch yet, more specifically, the "Prevent invalid userland pointer dereference" feature? I'm interested in rebuilding my production environment with this patch but I'm curious if anybody's noticed the performance penalties from using the patch or is it too minimal to even notice? thanks!