Re: Weird regression with -current after Jan 31st

[Tobias Nygren <tnn%NetBSD.org@localhost>]

On Sat, 20 Feb 2010 11:13:33 -0800
Scott Ellis <scotte%intrepid.warped.com@localhost> wrote:

> I was doing typical testing of a -current kernel with a 
> less-than-current userland, before installing all of userland, and ran 
> into a weird regression problem.

Looks like the same issue I'm seeing.

# sysctl -w vm.user_va0_disable=0

helps, but I wonder why some binaries are mapping stuff at page zero.
Also, we need a more descriptive error message than "Abort".

-Tobias