Re: sparc64 vs. gcc 4.8

[Michael <macallan%netbsd.org@localhost>]

Hello,

On Fri, 3 Oct 2014 16:15:39 +0000 (UTC)
Eduardo Horvath <eeh%NetBSD.org@localhost> wrote:

> One thing to look at is the memory synchronization/locking routines.
> I was looking at running the kernel in RMO mode and found that the
> memory barriers used in the mutex code are just... wrong.

I've been wondering if gcc 4.8 isn't really at fault here and just
exposes / no longer hides a bug or ten in our code, especially since
userland code is fine and we're seeing similar problems on powerpc.

have fun
Michael