Re: CVS commit: src/sys/arch

[Maxime Villard <max%m00nbsd.net@localhost>]

Le 23/02/2018 à 13:01, Joerg Sonnenberger a écrit :
> On Fri, Feb 23, 2018 at 09:00:56AM +0000, Maxime Villard wrote:
> > Module Name:	src
> > Committed By:	maxv
> > Date:		Fri Feb 23 09:00:56 UTC 2018
> >
> > Modified Files:
> > 	src/sys/arch/amd64/conf: Makefile.amd64
> > 	src/sys/arch/i386/conf: Makefile.i386
> >
> > Log Message:
> > Add -fno-shrink-wrap, to force GCC to push the frames at the very beginning
> > of the functions. Otherwise DDB is unable to display a correct stack trace
> > if a fault occurred in a function before the frame was pushed.
> >
> > Discussed on tech-kern@, flag suggested by Krister Walfridsson. Should fix
> > PR/52560.
>
> Please revert this. This makes the mistake of unconditionally adding
> -fno-omit-frame-pointer even worse as well as breaking the build with
> clang unconditionally.

Ok, I'll look for the Clang equivalent.

> Heck, it wasn't even discussed on tech-kern.

It was discussed on tech-kern.

> If anything, this should be done in the kernel config next to the
> -fno-omit-frame-potiner setting. This is a concession to DDB only,
> unlike i.e. -mcmode=kernel, it is not necessary for ABI reasons or
> correctness,

Moving -fno-shrink-wrap next to -fno-omit-frame-pointer did not fix the
issues. That is to say, the frame was _not_ pushed at the beginning of the
functions.