On Mon, Apr 23, 2012 at 10:13:09AM +0200, Jean-Yves Migeon wrote: > Careful here; Some Other Operating Systems report data address fault > with si_addr + SIGBUS, not instruction fault.
Maybe, but we need to define what NetBSD should do and test for that. I have not checked what NetBSD/amd64 does (as your code implied it would be broken), but NetBSD/sparc64 certainly delivers the fault-causing instruction address in si_addr. Martin
