Re: [PATCH v9 01/17] h8300: Assembly headers.

[Tobias Klauser <tklauser@distanz.ch>]

On 2015-04-27 at 09:48:39 +0200, Arnd Bergmann <arnd@arndb.de> wrote:
> On Monday 27 April 2015 09:42:41 Tobias Klauser wrote:
> > > diff --git a/arch/h8300/include/asm/elf.h b/arch/h8300/include/asm/elf.h
> > > new file mode 100644
> > > index 0000000..09031d0
> > > --- /dev/null
> > > +++ b/arch/h8300/include/asm/elf.h
> > > @@ -0,0 +1,101 @@
> > > +#ifndef __ASM_H8300_ELF_H
> > > +#define __ASM_H8300_ELF_H
> > > +
> > > +/*
> > > + * ELF register definitions..
> > > + */
> > > +
> > > +#include <asm/ptrace.h>
> > > +#include <asm/user.h>
> > > +
> > > +typedef unsigned long elf_greg_t;
> > > +
> > > +#define ELF_NGREG (sizeof(struct user_regs_struct) / sizeof(elf_greg_t))
> > > +typedef elf_greg_t elf_gregset_t[ELF_NGREG];
> > > +typedef unsigned long elf_fpregset_t;
> > > +
> > > +/*
> > > + * This is used to ensure we don't load something for the wrong architecture.
> > > + */
> > > +#define elf_check_arch(x) ((x)->e_machine == EM_H8_300)
> > 
> > EM_H8_300 is still used before it is introduced in patch 15/17, please
> > change the patch order. Otherwise you break bisectability.
> 
> While that is true in principle, I really wouldn't care about that
> when introducing a new architecture: There is no way to use this
> code unless you introduce all code first.

Agreed. But should the ELF machine at least be introduced before the
build infrastructure (patch 10/17) is added? Otherwise we're able to
compile the new arch port in principle but it will fail due to the
missing definition.