Re: [git pull -tip] headers_check fixes for other architectures

[Ingo Molnar <mingo@elte.hu>]

* Jaswinder Singh Rajput <jaswinder@kernel.org> wrote:

> On Thu, 2009-02-05 at 20:19 +0100, Ingo Molnar wrote:
> 
> > Jaswinder, because they can break the build we should proactively drop all 
> > architecture patches that do asm/types.h conversions.
> > 
> > I did build all the affected architectures via their defconfigs and they 
> > built just fine - but i cannot do wide coverage testing of them.
> > 
> > So i think we should drop these bits:
> > 
> >  earth4:~/tip> gll --grep='asm/types.h' linus..core/header-fixes
> >  1ff8f73: headers_check fix: xtensa, swab.h
> >  4810987: headers_check fix: powerpc, swab.h
> >  9f2cd96: headers_check fix: powerpc, kvm.h
> >  785857f: headers_check fix: powerpc, elf.h
> >  4be2c7f: headers_check fix: powerpc, bootx.h
> >  726da1e: headers_check fix: parisc, swab.h
> >  bef53ca: headers_check fix: mn10300, swab.h
> >  a9f6acc: headers_check fix: mips, swab.h
> >  d8cbec1: headers_check fix: m32r, swab.h
> >  040c92b: headers_check fix: ia64, swab.h
> >  6ce7950: headers_check fix: ia64, kvm.h
> >  fa9ea6c: headers_check fix: ia64, fpu.h
> >  295803e: headers_check fix: h8300, swab.h
> >  dacd762: headers_check fix: frv, swab.h
> >  350eb8b: headers_check fix: blackfin, swab.h
> >  1c6ce70: headers_check fix: avr32, swab.h
> >  e42ec24: headers_check fix: arm, swab.h
> >  4af3bf6: headers_check fix: arm, setup.h
> >  f100e6d: headers_check fix: arm, a.out.h
> >  3fd5906: headers_check fix: alpha, swab.h
> > 
> > and send the rest to Linus if all outstanding observations have been 
> > addressed. What do you think?
> > 
> 
> The problem is if we include linux/types.h in assembly file we will get
> error.
> 
> We can solve this problem by four options (or may be more):
> 1. by wrapping #ifndef __ASSEMBLY__ files which are used by assembly
> code as suggested by Tony.
> 
> 2. fix linux/types.h so that assembly file can also use it
> 
> 3. drop patches which are used by assembly files
> 
> 4. drop all architecture patches as you suggested.

hm, the highest quality approach seems to be #2, right?

	Ingo