Use -ffreestanding?

[Adrian Bunk <bunk@stusta.de>]

On Mon, Nov 08, 2004 at 08:01:30PM +0100, Andi Kleen wrote:
> On Mon, Nov 08, 2004 at 07:34:49PM +0100, Adrian Bunk wrote:
> > On Mon, Nov 08, 2004 at 06:51:20PM +0100, Andi Kleen wrote:
> > > On Mon, Nov 08, 2004 at 05:31:01PM +0100, Adrian Bunk wrote:
> > > > On Mon, Nov 08, 2004 at 05:19:35PM +0100, Andi Kleen wrote:
> > > > > > Rethinking it, I don't even understand the sprintf example in your 
> > > > > > changelog entry - shouldn't an inclusion of kernel.h always get it 
> > > > > > right?
> > > > > 
> > > > > Newer gcc rewrites sprintf(buf,"%s",str) to strcpy(buf,str) transparently.
> > > > 
> > > > Which gcc is "Newer"?
> > > 
> > > I saw it with 3.3-hammer, which had additional optimizations in this 
> > > area at some point. Note that 3.3-hammer is widely used. I don't 
> > > know if 3.4 does it in the same way.
> > 
> > Is this a -hammer specific problem?
> 
> No, I just checked a 4.0 mainline gcc and it does it too.
> 
> Note I saw it on x86-64, don't know if it occurs on i386 too.

OK, I see the difference:
After removing -fno-unit-at-a-time, I see this problem, too.

Why doesn't the kernel use -ffreestanding which should prevent all such 
problems?

> -Andi

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed