On Mon, Oct 5, 2015 at 8:25 AM, Michael Catanzaro mcatanzaro@gnome.org wrote:
On Mon, 2015-10-05 at 07:02 -0700, Andrew Lutomirski wrote:
Why is the filter causing SIGSYS instead of forcing an ENOSYS return?
I'll look into the abrt thing. It might be an easy fix.
--Andy
Simply because it's an experimental project, and it's much easier to crash with a core dump so it can be debugged, than to have obscure failures all over the place.
For deployment, though, I think the other approach is better. If, say, memfd_create returns ENOSYS, most libraries will fall back to older mechanisms.
--Andy