Re: recovering from SIGBUS in jitted code

  • From: Mike Pall <mike-1211@xxxxxxxxxx>
  • To: luajit@xxxxxxxxxxxxx
  • Date: Thu, 1 Nov 2012 00:34:57 +0100

aaron foo wrote:
> guess can do this by some how finding and jumping to an exit point
> in the currently executing trace`s MCode but then the state will be
> messed up anyway. maybe the cleanest fix would be to map something
> real to the address space then wait for the hook to get executed.

IMHO that doesn't look like a reasonable approach. Why don't you
check the access bounds instead? It's not that costly.

--Mike

Other related posts: