2006-09-26 21:03:18

by David Woodhouse

Subject: Re: [PATCH] avr32 architecture

On Tue, 2006-09-26 at 16:01 +0000, Linux Kernel Mailing List wrote:
> [PATCH] avr32 architecture

pmac /pmac/git/linux-2.6 $ make ARCH=avr32 headers_check
CHK include/linux/version.h
make[1]: `scripts/unifdef' is up to date.
CHECK include/asm/user.h
CHECK include/asm/unistd.h
/pmac/git/linux-2.6/usr/include/asm/unistd.h requires linux/linkage.h, which does not exist in exported headers
make[2]: *** [/pmac/git/linux-2.6/usr/include/asm/.check.unistd.h] Error 1

--
dwmw2

2006-09-26 21:17:31

by Håvard Skinnemoen

Subject: Re: [PATCH] avr32 architecture

On 9/26/06, David Woodhouse <[email protected]> wrote:
> On Tue, 2006-09-26 at 16:01 +0000, Linux Kernel Mailing List wrote:
> > [PATCH] avr32 architecture
>
> pmac /pmac/git/linux-2.6 $ make ARCH=avr32 headers_check
> CHK include/linux/version.h
> make[1]: `scripts/unifdef' is up to date.
> CHECK include/asm/user.h
> CHECK include/asm/unistd.h
> /pmac/git/linux-2.6/usr/include/asm/unistd.h requires linux/linkage.h, which does not exist in exported headers

Right. avr32-implement-kernel_execve.patch, which was flagged as "Will
merge" by Andrew, fixes it by moving the #ifdef __KERNEL__ guard to
cover everything but the __NR_foo definitions. So as long as that
one's still scheduled for inclusion, I think sending a separate patch
to fix this problem will do more harm than good.

But thanks for testing :)

H?vard

2006-09-26 23:41:00

by Andrew Morton

Subject: Re: [PATCH] avr32 architecture

On Tue, 26 Sep 2006 23:17:29 +0200
"Haavard Skinnemoen" <[email protected]> wrote:

> On 9/26/06, David Woodhouse <[email protected]> wrote:
> > On Tue, 2006-09-26 at 16:01 +0000, Linux Kernel Mailing List wrote:
> > > [PATCH] avr32 architecture
> >
> > pmac /pmac/git/linux-2.6 $ make ARCH=avr32 headers_check
> > CHK include/linux/version.h
> > make[1]: `scripts/unifdef' is up to date.
> > CHECK include/asm/user.h
> > CHECK include/asm/unistd.h
> > /pmac/git/linux-2.6/usr/include/asm/unistd.h requires linux/linkage.h, which does not exist in exported headers
>
> Right. avr32-implement-kernel_execve.patch, which was flagged as "Will
> merge" by Andrew, fixes it by moving the #ifdef __KERNEL__ guard to
> cover everything but the __NR_foo definitions.

Ho hum, such leakage happens sometimes.

> So as long as that
> one's still scheduled for inclusion, I think sending a separate patch
> to fix this problem will do more harm than good.
>

Yes, I'm planning on merging the execve cleanups for 2.6.19. It's probably
a week away yet.