Hi James / Fengguang,
On Mon, May 14, 2018 at 10:26:34AM +0100, James Hogan wrote:
> Hmm, I thought I'd fixed this, but again the kbuild test robot
> complains:
>
> On Sat, May 12, 2018 at 09:47:40AM +0800, kbuild test robot wrote:
> > bisected to: aea47daf8a396e512e0cfe11d9c05798749db172 compiler.h: Allow arch-specific overrides
> > commit date: 2 days ago
> > config: um-x86_64_defconfig (attached as .config)
> > compiler: gcc-7 (Debian 7.3.0-16) 7.3.0
> > reproduce:
> > git checkout aea47daf8a396e512e0cfe11d9c05798749db172
> > # save the attached .config to linux build tree
> > make ARCH=um SUBARCH=x86_64
> >
> > All errors (new ones prefixed by >>):
> >
> > In file included from arch/um/include/shared/init.h:44:0,
> > from arch/um/kernel/config.c:8:
> > >> include/linux/compiler_types.h:58:10: fatal error: asm/compiler.h: No such file or directory
> > #include <asm/compiler.h>
> > ^~~~~~~~~~~~~~~~
> > compilation terminated.
> >
> > vim +58 include/linux/compiler_types.h
> >
> > 56
> > 57 /* Allow architectures to override some definitions where necessary */
> > > 58 #include <asm/compiler.h>
> > 59
>
> Can anybody else reproduce that or have ideas why its still happening? I
> don't seem to be able to.
>
> Its from my mips-next-test branch here (that isn't in linux-next):
> git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips.git
I think I figured it out at last - it's because the generated includes
will be in $(objtree) but your patch looked for them in $(srctree).
Fixed in the v6 I just submitted.
Presumably the kbuild test robot is building with O=somewhere, but it
doesn't indicate that in the email which means the command it reports
will reproduce the bug simply doesn't.
I realise this is probably rare in the grand scheme of things, but maybe
it'd be worth mentioning O= in the emails Fengguang?
Thanks,
Paul
