Le 30/07/2016 à 19:35, Thor Lancelot Simon a écrit :
On Sat, Jul 30, 2016 at 07:12:31PM +0200, Joerg Sonnenberger wrote:
For what it is worth, I do believe that the handling of the 0 page
should be part of UVM and not pmap. I am only objection to forcing it
unconditionally.
Perhaps I misunderstood, but I thought we were all already agreed that
conditionalizing this upon a sysctl that could be set only when the
system is at insecure runlevel was an acceptable thing to do. I
mean, that's *one read of an integer* per check; surely that must
be acceptable -- Max?
My first proposion did not take care of the sysctl, because I didn't think it
was relevant. My last one does. Here is a patch for it. I believe it is safe
then to remove the va0_disable check in uvm_map.
http://m00nbsd.net/garbage/uvm/execmin.diff
