[PATCH 0/6] kill i386 and x86_64 directories

[PATCH 0/6] kill i386 and x86_64 directories

[Sam Ravnborg]

This series kill the old i386 and x86_64 directories.
The relevant files are moved and adapted and
Kconfig.debug was consolidated (thanks to Randy).

I had to modify both the top-lvel Makefile and the kconfig
Makefile to accomplish this. It was done in such a way that
it is trivial for other archs to use the same mechanism
should they have the need.

To solve the defconfig issue (i386 and x86_64 cannot share one)
the arch/x86/configs/ directory were introduced. This has
been used by other archs for some time now but x86 had not had
the need until now.

The Kconfig files could be consolidated much more - I only
did the minimal changes to make it work in the new place.

Total diffstat (in git format):

 Makefile                                           |    7 ++-
 arch/i386/.gitignore                               |    1 -
 arch/{i386 => x86}/Kconfig.cpu                     |    0 
 arch/{i386 => x86}/Kconfig.debug                   |   51 +++++++++++++---
 arch/{i386/Kconfig => x86/Kconfig.i386}            |    4 +-
 arch/{x86_64/Kconfig => x86/Kconfig.x86_64}        |    6 +--
 arch/x86/Makefile                                  |   16 +++++
 arch/{i386/Makefile => x86/Makefile_32}            |    7 +--
 arch/{i386/Makefile.cpu => x86/Makefile_32.cpu}    |    0 
 arch/{x86_64/Makefile => x86/Makefile_64}          |    5 +-
 .../{i386/defconfig => x86/configs/i386_defconfig} |    0 
 .../defconfig => x86/configs/x86_64_defconfig}     |    0 
 arch/x86_64/.gitignore                             |    1 -
 arch/x86_64/Kconfig.debug                          |   61 --------------------
 scripts/kconfig/Makefile                           |   35 +++++++-----
 15 files changed, 89 insertions(+), 105 deletions(-)

Patches follows.
The full serie can be pulled from:

	git://git.kernel.org/pub/scm/linux/kernel/git/sam/x86.git

Note: The serie were slightly modfied compared to the fist published
      version.

	Sam

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Sam Ravnborg]

On Thu, Oct 25, 2007 at 09:56:44PM +0200, Sam Ravnborg wrote:
> This series kill the old i386 and x86_64 directories.
> The relevant files are moved and adapted and
> Kconfig.debug was consolidated (thanks to Randy).
> 
> I had to modify both the top-lvel Makefile and the kconfig
> Makefile to accomplish this. It was done in such a way that
> it is trivial for other archs to use the same mechanism
> should they have the need.
> 
> To solve the defconfig issue (i386 and x86_64 cannot share one)
> the arch/x86/configs/ directory were introduced. This has
> been used by other archs for some time now but x86 had not had
> the need until now.
> 
> The Kconfig files could be consolidated much more - I only
> did the minimal changes to make it work in the new place.
> 
> Total diffstat (in git format):
> 
>  Makefile                                           |    7 ++-
>  arch/i386/.gitignore                               |    1 -
>  arch/{i386 => x86}/Kconfig.cpu                     |    0 
>  arch/{i386 => x86}/Kconfig.debug                   |   51 +++++++++++++---
>  arch/{i386/Kconfig => x86/Kconfig.i386}            |    4 +-
>  arch/{x86_64/Kconfig => x86/Kconfig.x86_64}        |    6 +--
>  arch/x86/Makefile                                  |   16 +++++
>  arch/{i386/Makefile => x86/Makefile_32}            |    7 +--
>  arch/{i386/Makefile.cpu => x86/Makefile_32.cpu}    |    0 
>  arch/{x86_64/Makefile => x86/Makefile_64}          |    5 +-
>  .../{i386/defconfig => x86/configs/i386_defconfig} |    0 
>  .../defconfig => x86/configs/x86_64_defconfig}     |    0 
>  arch/x86_64/.gitignore                             |    1 -
>  arch/x86_64/Kconfig.debug                          |   61 --------------------
>  scripts/kconfig/Makefile                           |   35 +++++++-----
>  15 files changed, 89 insertions(+), 105 deletions(-)
> 

By the way I consider this -rc1 material.
Because:
1) It sort of finishes the initial merge
2) Any breakage should be easy to trigger (build breakage) and easy to fix.

It is 1) that make me say this is -rc1 materail,
 and 2) that say that this is an acceptable 'breaking the rules' patch serie.

	Sam

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Thomas Gleixner]

Sam,

On Thu, 25 Oct 2007, Sam Ravnborg wrote:
> This series kill the old i386 and x86_64 directories.
> The relevant files are moved and adapted and
> Kconfig.debug was consolidated (thanks to Randy).
> 
> I had to modify both the top-lvel Makefile and the kconfig
> Makefile to accomplish this. It was done in such a way that
> it is trivial for other archs to use the same mechanism
> should they have the need.
> 
> To solve the defconfig issue (i386 and x86_64 cannot share one)
> the arch/x86/configs/ directory were introduced. This has
> been used by other archs for some time now but x86 had not had
> the need until now.
> 
> The Kconfig files could be consolidated much more - I only
> did the minimal changes to make it work in the new place.

Thanks for doing this.

I think the last remaining bit to cleanup is the symlink from
arch/x86/boot/bzImage. Now that the old directories are gone it does
not make much sense to keep them alive. Andrew will survive it :)

Please send it Linus wards and feel free to add my Acked-by to all of
them.

Thanks,

	tglx

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Kay Sievers]

On 10/25/07, Sam Ravnborg <sam@ravnborg.org> wrote:
> On Thu, Oct 25, 2007 at 09:56:44PM +0200, Sam Ravnborg wrote:
> > This series kill the old i386 and x86_64 directories.
> > The relevant files are moved and adapted and
> > Kconfig.debug was consolidated (thanks to Randy).
> >
> > I had to modify both the top-lvel Makefile and the kconfig
> > Makefile to accomplish this. It was done in such a way that
> > it is trivial for other archs to use the same mechanism
> > should they have the need.
> >
> > To solve the defconfig issue (i386 and x86_64 cannot share one)
> > the arch/x86/configs/ directory were introduced. This has
> > been used by other archs for some time now but x86 had not had
> > the need until now.
> >
> > The Kconfig files could be consolidated much more - I only
> > did the minimal changes to make it work in the new place.
> >
> > Total diffstat (in git format):
> >
> >  Makefile                                           |    7 ++-
> >  arch/i386/.gitignore                               |    1 -
> >  arch/{i386 => x86}/Kconfig.cpu                     |    0
> >  arch/{i386 => x86}/Kconfig.debug                   |   51 +++++++++++++---
> >  arch/{i386/Kconfig => x86/Kconfig.i386}            |    4 +-
> >  arch/{x86_64/Kconfig => x86/Kconfig.x86_64}        |    6 +--
> >  arch/x86/Makefile                                  |   16 +++++
> >  arch/{i386/Makefile => x86/Makefile_32}            |    7 +--
> >  arch/{i386/Makefile.cpu => x86/Makefile_32.cpu}    |    0
> >  arch/{x86_64/Makefile => x86/Makefile_64}          |    5 +-
> >  .../{i386/defconfig => x86/configs/i386_defconfig} |    0
> >  .../defconfig => x86/configs/x86_64_defconfig}     |    0
> >  arch/x86_64/.gitignore                             |    1 -
> >  arch/x86_64/Kconfig.debug                          |   61 --------------------
> >  scripts/kconfig/Makefile                           |   35 +++++++-----
> >  15 files changed, 89 insertions(+), 105 deletions(-)
> >
>
> By the way I consider this -rc1 material.
> Because:
> 1) It sort of finishes the initial merge
> 2) Any breakage should be easy to trigger (build breakage) and easy to fix.
>
> It is 1) that make me say this is -rc1 materail,
>  and 2) that say that this is an acceptable 'breaking the rules' patch serie.

Hey Sam,
I get this after a simple pull from your tree and a: make clean; make:
  LD      init/built-in.o
  LD      .tmp_vmlinux1
arch/x86/kernel/head_64.o: In function `early_idt_handler':
(.text.head+0x1e5): undefined reference to `early_printk'
arch/x86/kernel/head64.o: In function `x86_64_start_kernel':
head64.c:(.init.text+0xa8): undefined reference to `early_printk'
arch/x86/kernel/built-in.o: In function `early_panic':
(.text+0x9c07): undefined reference to `early_printk'
arch/x86/mm/built-in.o: In function `init_memory_mapping':
(.init.text+0xc0a): undefined reference to `early_printk'
make: *** [.tmp_vmlinux1] Error 1

Cheers,
Kay

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Sam Ravnborg]

On Thu, Oct 25, 2007 at 10:19:40PM +0200, Kay Sievers wrote:
> > Because:
> > 1) It sort of finishes the initial merge
> > 2) Any breakage should be easy to trigger (build breakage) and easy to fix.
> >
> > It is 1) that make me say this is -rc1 materail,
> >  and 2) that say that this is an acceptable 'breaking the rules' patch serie.
> 
> Hey Sam,
> I get this after a simple pull from your tree and a: make clean; make:
>   LD      init/built-in.o
>   LD      .tmp_vmlinux1
> arch/x86/kernel/head_64.o: In function `early_idt_handler':
> (.text.head+0x1e5): undefined reference to `early_printk'
> arch/x86/kernel/head64.o: In function `x86_64_start_kernel':
> head64.c:(.init.text+0xa8): undefined reference to `early_printk'
> arch/x86/kernel/built-in.o: In function `early_panic':
> (.text+0x9c07): undefined reference to `early_printk'
> arch/x86/mm/built-in.o: In function `init_memory_mapping':
> (.init.text+0xc0a): undefined reference to `early_printk'
> make: *** [.tmp_vmlinux1] Error 1

I think this is my changes to the Kconfig.debug path
I received from Randy.

Checking....
Yup - reverting my change bring back EARLY_PRINTK wich is
always required on x86_64.

I will redo the series - and build test a bit more.
Thanks.

	Sam

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Bodo Eggert]

Thomas Gleixner <tglx@linutronix.de> wrote:

> I think the last remaining bit to cleanup is the symlink from
> arch/x86/boot/bzImage.

BTW: Is it useful to have (b)zimage under $ARCH while vmlinux is in the root
dir? (Besides being compatible to external scripts)
-- 
I always tell customers/clients the same thing:
   "Good, Fast, Cheap.  You can pick two."
        -- randem in <slrna09rui.g43.root@jade.randemmedia.com>
Friß, Spammer: 38Ewdi@3.7eggert.dyndns.org mQIC5mdv@t8x8ycg.7eggert.dyndns.org

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Sam Ravnborg]

On Thu, Oct 25, 2007 at 10:17:53PM +0200, Thomas Gleixner wrote:
> 
> I think the last remaining bit to cleanup is the symlink from
> arch/x86/boot/bzImage. Now that the old directories are gone it does
> not make much sense to keep them alive. Andrew will survive it :)

It is not just Andrew I'm afraid.
We got a Debian bug-report too when we did not have the symlink and I
asume this is true for several others.
We should let it stay in 2.6.24 and remove the symlink for 2.6.25 IMO.

	Sam

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Alan Cox]

On Fri, 26 Oct 2007 08:23:36 +0200
Sam Ravnborg <sam@ravnborg.org> wrote:

> On Thu, Oct 25, 2007 at 10:17:53PM +0200, Thomas Gleixner wrote:
> > 
> > I think the last remaining bit to cleanup is the symlink from
> > arch/x86/boot/bzImage. Now that the old directories are gone it does
> > not make much sense to keep them alive. Andrew will survive it :)
> 
> It is not just Andrew I'm afraid.
> We got a Debian bug-report too when we did not have the symlink and I
> asume this is true for several others.
> We should let it stay in 2.6.24 and remove the symlink for 2.6.25 IMO.

I would put it on the list and leave it for a couple of years so that
people using old distros have all upgraded at that point. Adrian will
remember to delete it if its on the list to go ....

Re: [PATCH 0/6] kill i386 and x86_64 directories

[Arjan van de Ven]

On Fri, 26 Oct 2007 08:23:36 +0200
Sam Ravnborg <sam@ravnborg.org> wrote:

> On Thu, Oct 25, 2007 at 10:17:53PM +0200, Thomas Gleixner wrote:
> > 
> > I think the last remaining bit to cleanup is the symlink from
> > arch/x86/boot/bzImage. Now that the old directories are gone it does
> > not make much sense to keep them alive. Andrew will survive it :)
> 
> It is not just Andrew I'm afraid.
> We got a Debian bug-report too when we did not have the symlink and I
> asume this is true for several others.
> We should let it stay in 2.6.24 and remove the symlink for 2.6.25 IMO.

that's the worst of all things to do... either make a change in ONE go,
when you do all the changes anyway (users will remmeber 2.6.24 was the
unification kernel, but 2.6.25 breaking stuff will be a total
surprise)... or keep them around forever. As long as it works, nobody
will change the scripts.