architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Adrian Bunk]

On Mon, Jun 16, 2008 at 01:26:54PM +0100, David Woodhouse wrote:
>...
> Actually, I think we can drop the preceding patch too -- it's not
> necessary to fix the regression in 2.6.26, and dhowells suggests that we
> might still want to use ARCH_SUPPORTS_AOUT, by making BINFMT_AOUT depend
> on it instead of that hard-coded list of architectures.
> 
> If we do that, then ARCH_SUPPORTS_AOUT would need to be defined on a
> different set of architectures to the ones it's currently defined on --
> so I don't really want to go there before 2.6.26. I'll push just the
> parts which were included in the mail I just sent.
> 
> Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> sense?

My guess would be that binfmt_aout isn't actually used there.

The affected architectures are:
- h8300
- m32r
- parisc
- sh
- v850

v850 is broken beyond repair, and I hope my patch to remove the 
architecture will make it into 2.6.27.

For the other 4 architectures I have Cc'ed the architecture maintainers.

> dwmw2

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Tue, 2008-06-17 at 11:42 +0300, Adrian Bunk wrote:
> 
> > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > sense?
> 
> My guess would be that binfmt_aout isn't actually used there.
> 
> The affected architectures are:
> - h8300
> - m32r
> - parisc
> - sh
> - v850
> 
> v850 is broken beyond repair, and I hope my patch to remove the 
> architecture will make it into 2.6.27.
> 
> For the other 4 architectures I have Cc'ed the architecture maintainers.

There's another set of architectures which have <asm/a.out.h> but don't
define ARCH_SUPPORTS_AOUT. Including PowerPC.

I think we can probably get away with removing <asm/a.out.h> from every
architecture other than Alpha, ARM, m68k and x86, and killing
ARCH_SUPPORTS_AOUT completely.

Or maybe, as dhowells suggests, re-defining ARCH_SUPPORTS_AOUT to be set
only on Alpha, ARM, m68k and x86 (so BINFMT_AOUT can depend on it).

-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Adrian Bunk]

On Tue, Jun 17, 2008 at 10:46:44AM +0100, David Woodhouse wrote:
> On Tue, 2008-06-17 at 11:42 +0300, Adrian Bunk wrote:
> > 
> > > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > > sense?
> > 
> > My guess would be that binfmt_aout isn't actually used there.
> > 
> > The affected architectures are:
> > - h8300
> > - m32r
> > - parisc
> > - sh
> > - v850
> > 
> > v850 is broken beyond repair, and I hope my patch to remove the 
> > architecture will make it into 2.6.27.
> > 
> > For the other 4 architectures I have Cc'ed the architecture maintainers.
> 
> There's another set of architectures which have <asm/a.out.h> but don't
> define ARCH_SUPPORTS_AOUT. Including PowerPC.
> 
> I think we can probably get away with removing <asm/a.out.h> from every
> architecture other than Alpha, ARM, m68k and x86, and killing
> ARCH_SUPPORTS_AOUT completely.

include/linux/a.out.h contains an #include <asm/a.out.h> and is exported 
to userspace...

> Or maybe, as dhowells suggests, re-defining ARCH_SUPPORTS_AOUT to be set
> only on Alpha, ARM, m68k and x86 (so BINFMT_AOUT can depend on it).

The pattern for this stuff is to provide a HAVE_AOUT kconfig variable 
that gets select'ed by these architectures.

> dwmw2

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Tue, 2008-06-17 at 13:11 +0300, Adrian Bunk wrote:
> On Tue, Jun 17, 2008 at 10:46:44AM +0100, David Woodhouse wrote:
> > On Tue, 2008-06-17 at 11:42 +0300, Adrian Bunk wrote:
> > > 
> > > > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > > > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > > > sense?
> > > 
> > > My guess would be that binfmt_aout isn't actually used there.
> > > 
> > > The affected architectures are:
> > > - h8300
> > > - m32r
> > > - parisc
> > > - sh
> > > - v850
> > > 
> > > v850 is broken beyond repair, and I hope my patch to remove the 
> > > architecture will make it into 2.6.27.
> > > 
> > > For the other 4 architectures I have Cc'ed the architecture maintainers.
> > 
> > There's another set of architectures which have <asm/a.out.h> but don't
> > define ARCH_SUPPORTS_AOUT. Including PowerPC.
> > 
> > I think we can probably get away with removing <asm/a.out.h> from every
> > architecture other than Alpha, ARM, m68k and x86, and killing
> > ARCH_SUPPORTS_AOUT completely.
> 
> include/linux/a.out.h contains an #include <asm/a.out.h> and is exported 
> to userspace...

Actually, it isn't -- but it _should_ be. But we can do it conditionally.
See http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;hY39c6a41

> > Or maybe, as dhowells suggests, re-defining ARCH_SUPPORTS_AOUT to be set
> > only on Alpha, ARM, m68k and x86 (so BINFMT_AOUT can depend on it).
> 
> The pattern for this stuff is to provide a HAVE_AOUT kconfig variable 
> that gets select'ed by these architectures.

Or that; it doesn't really matter. I was about to remove
ARCH_SUPPORTS_AOUT completely, but David suggested that it could be used
to kill that hard-coded '(X86_32 || ALPHA || ARM || M68K)' in
fs/Kconfig.binfmt.


-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Adrian Bunk]

On Tue, Jun 17, 2008 at 11:18:45AM +0100, David Woodhouse wrote:
> On Tue, 2008-06-17 at 13:11 +0300, Adrian Bunk wrote:
> > On Tue, Jun 17, 2008 at 10:46:44AM +0100, David Woodhouse wrote:
> > > On Tue, 2008-06-17 at 11:42 +0300, Adrian Bunk wrote:
> > > > 
> > > > > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > > > > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > > > > sense?
> > > > 
> > > > My guess would be that binfmt_aout isn't actually used there.
> > > > 
> > > > The affected architectures are:
> > > > - h8300
> > > > - m32r
> > > > - parisc
> > > > - sh
> > > > - v850
> > > > 
> > > > v850 is broken beyond repair, and I hope my patch to remove the 
> > > > architecture will make it into 2.6.27.
> > > > 
> > > > For the other 4 architectures I have Cc'ed the architecture maintainers.
> > > 
> > > There's another set of architectures which have <asm/a.out.h> but don't
> > > define ARCH_SUPPORTS_AOUT. Including PowerPC.
> > > 
> > > I think we can probably get away with removing <asm/a.out.h> from every
> > > architecture other than Alpha, ARM, m68k and x86, and killing
> > > ARCH_SUPPORTS_AOUT completely.
> > 
> > include/linux/a.out.h contains an #include <asm/a.out.h> and is exported 
> > to userspace...
> 
> Actually, it isn't -- but it _should_ be. But we can do it conditionally.
> See http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;hY39c6a41
>...

It's already in Linus' tree.

Sorry, you are right, I missed that it takes care of exporting 
linux/a.out.h only when asm/a.out.h exists.

> dwmw2

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Tue, 2008-06-17 at 13:11 +0300, Adrian Bunk wrote:
> The pattern for this stuff is to provide a HAVE_AOUT kconfig variable 
> that gets select'ed by these architectures.

http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;h¥db5179

Subject: [PATCH] Introduce HAVE_AOUT symbol to remove hard-coded arch list for BINFMT_AOUT

HAVE_AOUT doesn't quite do the same thing as the recently removed
ARCH_SUPPORTS_AOUT config option. That was set even on platforms where
binfmt_aout isn't supported, although it's not entirely clear why.

So it's best just to introduce a new symbol, handled consistently with
other similar HAVE_xxx symbols; with a simple 'select' the arch Kconfig.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
---
 arch/alpha/Kconfig |    1 +
 arch/arm/Kconfig   |    1 +
 arch/m68k/Kconfig  |    1 +
 arch/x86/Kconfig   |    1 +
 fs/Kconfig.binfmt  |    5 ++++-
 5 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig
index 207196e..f80e548 100644
--- a/arch/alpha/Kconfig
+++ b/arch/alpha/Kconfig
@@ -7,6 +7,7 @@ config ALPHA
 	default y
 	select HAVE_IDE
 	select HAVE_OPROFILE
+	select HAVE_AOUT
 	help
 	  The Alpha is a 64-bit general-purpose processor designed and
 	  marketed by the Digital Equipment Corporation of blessed memory,
diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
index 8fd3ab3..2be8bd7 100644
--- a/arch/arm/Kconfig
+++ b/arch/arm/Kconfig
@@ -14,6 +14,7 @@ config ARM
 	select HAVE_OPROFILE
 	select HAVE_KPROBES if (!XIP_KERNEL)
 	select HAVE_KRETPROBES if (HAVE_KPROBES)
+	select HAVE_AOUT
 	help
 	  The ARM series is a line of low-power-consumption RISC chip designs
 	  licensed by ARM Ltd and targeted at embedded applications and
diff --git a/arch/m68k/Kconfig b/arch/m68k/Kconfig
index bfb6395..83c4338 100644
--- a/arch/m68k/Kconfig
+++ b/arch/m68k/Kconfig
@@ -6,6 +6,7 @@ config M68K
 	bool
 	default y
 	select HAVE_IDE
+	select HAVE_AOUT
 
 config MMU
 	bool
diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index 8fdaa70..bd066ae 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -25,6 +25,7 @@ config X86
 	select HAVE_KRETPROBES
 	select HAVE_KVM if ((X86_32 && !X86_VOYAGER && !X86_VISWS && !X86_NUMAQ) || X86_64)
 	select HAVE_ARCH_KGDB if !X86_VOYAGER
+	select HAVE_AOUT if X86_32
 
 config ARCH_DEFCONFIG
 	string
diff --git a/fs/Kconfig.binfmt b/fs/Kconfig.binfmt
index 0999d5c..5592e08 100644
--- a/fs/Kconfig.binfmt
+++ b/fs/Kconfig.binfmt
@@ -59,9 +59,12 @@ config BINFMT_SHARED_FLAT
 	help
 	  Support FLAT shared libraries
 
+config HAVE_AOUT
+       def_bool n
+
 config BINFMT_AOUT
 	tristate "Kernel support for a.out and ECOFF binaries"
-	depends on (X86_32 || ALPHA || ARM || M68K)
+	depends on HAVE_AOUT
 	---help---
 	  A.out (Assembler.OUTput) is a set of formats for libraries and
 	  executables used in the earliest versions of UNIX.  Linux used
-- 
1.5.5.1


-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Sam Ravnborg]

On Tue, Jun 17, 2008 at 12:28:33PM +0100, David Woodhouse wrote:
> On Tue, 2008-06-17 at 13:11 +0300, Adrian Bunk wrote:
> > The pattern for this stuff is to provide a HAVE_AOUT kconfig variable 
> > that gets select'ed by these architectures.
> 
> http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;h¥db5179
> 
> Subject: [PATCH] Introduce HAVE_AOUT symbol to remove hard-coded arch list for BINFMT_AOUT
> 
> HAVE_AOUT doesn't quite do the same thing as the recently removed
> ARCH_SUPPORTS_AOUT config option. That was set even on platforms where
> binfmt_aout isn't supported, although it's not entirely clear why.
> 
> So it's best just to introduce a new symbol, handled consistently with
> other similar HAVE_xxx symbols; with a simple 'select' the arch Kconfig.

Looks good.
One minor issue...


> @@ -7,6 +7,7 @@ config ALPHA
>  	default y
>  	select HAVE_IDE
>  	select HAVE_OPROFILE
> +	select HAVE_AOUT
Sorting these alphabetically will result in less merge conflicts when the list grows.

	Sam

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Tue, 2008-06-17 at 14:23 +0200, Sam Ravnborg wrote:
> Sorting these alphabetically will result in less merge conflicts when
> the list grows.

Maybe, but it doesn't grow that fast -- and it's small enough that if I
add stuff in the middle, the context will cover the whole thing and
cause a merge conflict either way. And it isn't in alphabetical order
now either :)

-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Sam Ravnborg]

On Tue, Jun 17, 2008 at 01:29:09PM +0100, David Woodhouse wrote:
> On Tue, 2008-06-17 at 14:23 +0200, Sam Ravnborg wrote:
> > Sorting these alphabetically will result in less merge conflicts when
> > the list grows.
> 
> Maybe, but it doesn't grow that fast -- and it's small enough that if I
> add stuff in the middle, the context will cover the whole thing and
> cause a merge conflict either way. And it isn't in alphabetical order
> now either :)
I know and I know.
If someone did the janitorial effort and converted all the
per-arch defined symbols to one single definiton
and select the list would grow rapidly.

	Sam

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Tue, 2008-06-17 at 14:37 +0200, Sam Ravnborg wrote:
> I know and I know.
> If someone did the janitorial effort and converted all the
> per-arch defined symbols to one single definiton
> and select the list would grow rapidly.

http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;hç5fd00c

They're still not in alphabetical order, but I haven't made it worse :)

-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Adrian Bunk]

On Tue, Jun 17, 2008 at 12:28:33PM +0100, David Woodhouse wrote:
> On Tue, 2008-06-17 at 13:11 +0300, Adrian Bunk wrote:
> > The pattern for this stuff is to provide a HAVE_AOUT kconfig variable 
> > that gets select'ed by these architectures.
> 
> http://git.infradead.org/users/dwmw2/aout-2.6.git?a=commitdiff;h¥db5179
> 
> Subject: [PATCH] Introduce HAVE_AOUT symbol to remove hard-coded arch list for BINFMT_AOUT
> 
> HAVE_AOUT doesn't quite do the same thing as the recently removed
> ARCH_SUPPORTS_AOUT config option. That was set even on platforms where
> binfmt_aout isn't supported, although it's not entirely clear why.
> 
> So it's best just to introduce a new symbol, handled consistently with
> other similar HAVE_xxx symbols; with a simple 'select' the arch Kconfig.
> 
> Signed-off-by: David Woodhouse <dwmw2@infradead.org>
> ---
>  arch/alpha/Kconfig |    1 +
>  arch/arm/Kconfig   |    1 +
>  arch/m68k/Kconfig  |    1 +
>  arch/x86/Kconfig   |    1 +
>  fs/Kconfig.binfmt  |    5 ++++-
>  5 files changed, 8 insertions(+), 1 deletions(-)
>...
> --- a/fs/Kconfig.binfmt
> +++ b/fs/Kconfig.binfmt
> @@ -59,9 +59,12 @@ config BINFMT_SHARED_FLAT
>  	help
>  	  Support FLAT shared libraries
>  
> +config HAVE_AOUT
> +       def_bool n
> +
>  config BINFMT_AOUT
>  	tristate "Kernel support for a.out and ECOFF binaries"
> -	depends on (X86_32 || ALPHA || ARM || M68K)
> +	depends on HAVE_AOUT
>  	---help---
>  	  A.out (Assembler.OUTput) is a set of formats for libraries and
>  	  executables used in the earliest versions of UNIX.  Linux used

What about m68knommu?

I don't know whether it was intentionally available there, but the 
removal in your patch does not seem to be intentional.

> dwmw2

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David Woodhouse]

On Sun, 2008-06-22 at 12:49 +0300, Adrian Bunk wrote:
> 
> What about m68knommu?
> 
> I don't know whether it was intentionally available there, but the 
> removal in your patch does not seem to be intentional.

Hm, true. Well spotted. I think it's probably correct though -- I don't
think m68knommu can really support a.out binaries, can it?

-- 
dwmw2

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[David McCullough]

Jivin David Woodhouse lays it down ...
> On Sun, 2008-06-22 at 12:49 +0300, Adrian Bunk wrote:
> > 
> > What about m68knommu?
> > 
> > I don't know whether it was intentionally available there, but the 
> > removal in your patch does not seem to be intentional.
> 
> Hm, true. Well spotted. I think it's probably correct though -- I don't
> think m68knommu can really support a.out binaries, can it?

No it can't.

-- 
David McCullough,  david_mccullough@securecomputing.com,   Ph:+61 734352815
Secure Computing - SnapGear  http://www.uCdot.org   http://www.snapgear.com

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Adrian Bunk]

On Sun, Jun 22, 2008 at 12:40:50PM +0100, David Woodhouse wrote:
> On Sun, 2008-06-22 at 12:49 +0300, Adrian Bunk wrote:
> > 
> > What about m68knommu?
> > 
> > I don't know whether it was intentionally available there, but the 
> > removal in your patch does not seem to be intentional.
> 
> Hm, true. Well spotted. I think it's probably correct though -- I don't
> think m68knommu can really support a.out binaries, can it?

I found another similar case:

X86_32 UML.

a.out is most likely working there?

> dwmw2

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Jeff Dike]

On Tue, Jun 24, 2008 at 02:46:43PM +0300, Adrian Bunk wrote:
> I found another similar case:
> 
> X86_32 UML.
> 
> a.out is most likely working there?

Yup, if it works on the host, it'll work in UML.

     	   	 Jeff

-- 
Work email - jdike at linux dot intel dot com

[2.6 patch] remove unused asm/a.out.h files

[Adrian Bunk]

On Tue, Jun 17, 2008 at 10:46:44AM +0100, David Woodhouse wrote:
>...
> I think we can probably get away with removing <asm/a.out.h> from every
> architecture other than Alpha, ARM, m68k and x86,
>...

Patch below.

> dwmw2

cu
Adrian


<--  snip  -->


This patch removes asm/a.out.h from all architectures that do not
provide a.out support.

This patch also includes the required removal of (unused)
#include <asm/a.out.h>'s and #include <linux/a.out.h>'s in the arch/
code for these architectures.

The ppc and v850 architectures that should anyway be removed in 2.6.27 
are not touched by this patch to avoid needless merge conflicts.

Signed-off-by: Adrian Bunk <bunk@kernel.org>

---

 arch/cris/arch-v10/boot/tools/build.c   |    1 
 arch/h8300/kernel/process.c             |    1 
 arch/ia64/mm/init.c                     |    1 
 arch/m68knommu/kernel/process.c         |    1 
 arch/m68knommu/kernel/traps.c           |    1 
 arch/mips/kernel/irixelf.c              |    1 
 arch/mips/kernel/process.c              |    1 
 arch/mips/kernel/syscall.c              |    1 
 arch/powerpc/kernel/softemu8xx.c        |    1 
 arch/powerpc/kernel/traps.c             |    1 
 arch/powerpc/platforms/chrp/setup.c     |    1 
 arch/powerpc/platforms/maple/setup.c    |    1 
 arch/powerpc/platforms/powermac/setup.c |    1 
 arch/powerpc/platforms/pseries/setup.c  |    1 
 include/asm-avr32/a.out.h               |   20 -------------
 include/asm-blackfin/a.out.h            |   19 -------------
 include/asm-cris/a.out.h                |   26 -----------------
 include/asm-h8300/a.out.h               |   20 -------------
 include/asm-ia64/a.out.h                |   32 ---------------------
 include/asm-m32r/a.out.h                |   20 -------------
 include/asm-m68knommu/a.out.h           |    1 
 include/asm-mips/a.out.h                |   35 ------------------------
 include/asm-parisc/a.out.h              |   20 -------------
 include/asm-powerpc/a.out.h             |   20 -------------
 include/asm-sh/a.out.h                  |   20 -------------
 include/asm-xtensa/a.out.h              |   29 -------------------
 26 files changed, 276 deletions(-)

9cd72b780be8d25cd2deb99a923030f2fe2aeba4 diff --git a/arch/cris/arch-v10/boot/tools/build.c b/arch/cris/arch-v10/boot/tools/build.c
index 2f9bbb2..c8adef3 100644
--- a/arch/cris/arch-v10/boot/tools/build.c
+++ b/arch/cris/arch-v10/boot/tools/build.c
@@ -30,7 +30,6 @@
 #include <sys/sysmacros.h>
 #include <unistd.h>	/* contains read/write */
 #include <fcntl.h>
-#include <linux/a.out.h>
 #include <errno.h>
 
 #define MINIX_HEADER 32
diff --git a/arch/h8300/kernel/process.c b/arch/h8300/kernel/process.c
index dfbe7ab..a8ef654 100644
--- a/arch/h8300/kernel/process.c
+++ b/arch/h8300/kernel/process.c
@@ -34,7 +34,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/interrupt.h>
 #include <linux/reboot.h>
 #include <linux/fs.h>
diff --git a/arch/ia64/mm/init.c b/arch/ia64/mm/init.c
index 200100e..f482a90 100644
--- a/arch/ia64/mm/init.c
+++ b/arch/ia64/mm/init.c
@@ -21,7 +21,6 @@
 #include <linux/bitops.h>
 #include <linux/kexec.h>
 
-#include <asm/a.out.h>
 #include <asm/dma.h>
 #include <asm/ia32.h>
 #include <asm/io.h>
diff --git a/arch/m68knommu/kernel/process.c b/arch/m68knommu/kernel/process.c
index 47502d5..3f2d774 100644
--- a/arch/m68knommu/kernel/process.c
+++ b/arch/m68knommu/kernel/process.c
@@ -25,7 +25,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/interrupt.h>
 #include <linux/reboot.h>
 #include <linux/fs.h>
diff --git a/arch/m68knommu/kernel/traps.c b/arch/m68knommu/kernel/traps.c
index ec9aea6..0157b9d 100644
--- a/arch/m68knommu/kernel/traps.c
+++ b/arch/m68knommu/kernel/traps.c
@@ -22,7 +22,6 @@
 #include <linux/mm.h>
 #include <linux/module.h>
 #include <linux/types.h>
-#include <linux/a.out.h>
 #include <linux/user.h>
 #include <linux/string.h>
 #include <linux/linkage.h>
diff --git a/arch/mips/kernel/irixelf.c b/arch/mips/kernel/irixelf.c
index 469c723..e068a72 100644
--- a/arch/mips/kernel/irixelf.c
+++ b/arch/mips/kernel/irixelf.c
@@ -18,7 +18,6 @@
 #include <linux/sched.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
-#include <linux/a.out.h>
 #include <linux/errno.h>
 #include <linux/init.h>
 #include <linux/signal.h>
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
index 2c09a44..2040fc5 100644
--- a/arch/mips/kernel/process.c
+++ b/arch/mips/kernel/process.c
@@ -22,7 +22,6 @@
 #include <linux/personality.h>
 #include <linux/sys.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/init.h>
 #include <linux/completion.h>
 #include <linux/kallsyms.h>
diff --git a/arch/mips/kernel/syscall.c b/arch/mips/kernel/syscall.c
index af1bdc8..0d2b609 100644
--- a/arch/mips/kernel/syscall.c
+++ b/arch/mips/kernel/syscall.c
@@ -7,7 +7,6 @@
  * Copyright (C) 1999, 2000 Silicon Graphics, Inc.
  * Copyright (C) 2001 MIPS Technologies, Inc.
  */
-#include <linux/a.out.h>
 #include <linux/capability.h>
 #include <linux/errno.h>
 #include <linux/linkage.h>
diff --git a/arch/powerpc/kernel/softemu8xx.c b/arch/powerpc/kernel/softemu8xx.c
index 67d6f68..2489493 100644
--- a/arch/powerpc/kernel/softemu8xx.c
+++ b/arch/powerpc/kernel/softemu8xx.c
@@ -23,7 +23,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/interrupt.h>
 
 #include <asm/pgtable.h>
diff --git a/arch/powerpc/kernel/traps.c b/arch/powerpc/kernel/traps.c
index 4b5b7ff..c7e499e 100644
--- a/arch/powerpc/kernel/traps.c
+++ b/arch/powerpc/kernel/traps.c
@@ -23,7 +23,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/interrupt.h>
 #include <linux/init.h>
 #include <linux/module.h>
diff --git a/arch/powerpc/platforms/chrp/setup.c b/arch/powerpc/platforms/chrp/setup.c
index 116babb..e6b3fac 100644
--- a/arch/powerpc/platforms/chrp/setup.c
+++ b/arch/powerpc/platforms/chrp/setup.c
@@ -17,7 +17,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/tty.h>
 #include <linux/major.h>
 #include <linux/interrupt.h>
diff --git a/arch/powerpc/platforms/maple/setup.c b/arch/powerpc/platforms/maple/setup.c
index 3647147..d4c61c3 100644
--- a/arch/powerpc/platforms/maple/setup.c
+++ b/arch/powerpc/platforms/maple/setup.c
@@ -23,7 +23,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/tty.h>
 #include <linux/string.h>
 #include <linux/delay.h>
diff --git a/arch/powerpc/platforms/powermac/setup.c b/arch/powerpc/platforms/powermac/setup.c
index 00bd016..19c287a 100644
--- a/arch/powerpc/platforms/powermac/setup.c
+++ b/arch/powerpc/platforms/powermac/setup.c
@@ -33,7 +33,6 @@
 #include <linux/ptrace.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/tty.h>
 #include <linux/string.h>
 #include <linux/delay.h>
diff --git a/arch/powerpc/platforms/pseries/setup.c b/arch/powerpc/platforms/pseries/setup.c
index f5d29f5..577efcd 100644
--- a/arch/powerpc/platforms/pseries/setup.c
+++ b/arch/powerpc/platforms/pseries/setup.c
@@ -25,7 +25,6 @@
 #include <linux/unistd.h>
 #include <linux/slab.h>
 #include <linux/user.h>
-#include <linux/a.out.h>
 #include <linux/tty.h>
 #include <linux/major.h>
 #include <linux/interrupt.h>
diff --git a/include/asm-avr32/a.out.h b/include/asm-avr32/a.out.h
deleted file mode 100644
index e46375a..0000000
--- a/include/asm-avr32/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef __ASM_AVR32_A_OUT_H
-#define __ASM_AVR32_A_OUT_H
-
-struct exec
-{
-  unsigned long a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* __ASM_AVR32_A_OUT_H */
diff --git a/include/asm-blackfin/a.out.h b/include/asm-blackfin/a.out.h
deleted file mode 100644
index 6c3d652..0000000
--- a/include/asm-blackfin/a.out.h
+++ /dev/null
@@ -1,19 +0,0 @@
-#ifndef __BFIN_A_OUT_H__
-#define __BFIN_A_OUT_H__
-
-struct exec {
-	unsigned long a_info;	/* Use macros N_MAGIC, etc for access */
-	unsigned a_text;	/* length of text, in bytes */
-	unsigned a_data;	/* length of data, in bytes */
-	unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-	unsigned a_syms;	/* length of symbol table data in file, in bytes */
-	unsigned a_entry;	/* start address */
-	unsigned a_trsize;	/* length of relocation info for text, in bytes */
-	unsigned a_drsize;	/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif				/* __BFIN_A_OUT_H__ */
diff --git a/include/asm-cris/a.out.h b/include/asm-cris/a.out.h
deleted file mode 100644
index c82e9f9..0000000
--- a/include/asm-cris/a.out.h
+++ /dev/null
@@ -1,26 +0,0 @@
-#ifndef __CRIS_A_OUT_H__
-#define __CRIS_A_OUT_H__
-
-/* we don't support a.out binaries on Linux/CRIS anyway, so this is
- * not really used but still needed because binfmt_elf.c for some reason
- * wants to know about a.out even if there is no interpreter available...
- */
-
-struct exec
-{
-  unsigned long a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif
diff --git a/include/asm-h8300/a.out.h b/include/asm-h8300/a.out.h
deleted file mode 100644
index ded780f..0000000
--- a/include/asm-h8300/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef __H8300_A_OUT_H__
-#define __H8300_A_OUT_H__
-
-struct exec
-{
-  unsigned long a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* __H8300_A_OUT_H__ */
diff --git a/include/asm-ia64/a.out.h b/include/asm-ia64/a.out.h
deleted file mode 100644
index 193dcfb..0000000
--- a/include/asm-ia64/a.out.h
+++ /dev/null
@@ -1,32 +0,0 @@
-#ifndef _ASM_IA64_A_OUT_H
-#define _ASM_IA64_A_OUT_H
-
-/*
- * No a.out format has been (or should be) defined so this file is
- * just a dummy that allows us to get binfmt_elf compiled.  It
- * probably would be better to clean up binfmt_elf.c so it does not
- * necessarily depend on there being a.out support.
- *
- * Modified 1998-2002
- *	David Mosberger-Tang <davidm@hpl.hp.com>, Hewlett-Packard Co.
- */
-
-#include <linux/types.h>
-
-struct exec {
-	unsigned long a_info;
-	unsigned long a_text;
-	unsigned long a_data;
-	unsigned long a_bss;
-	unsigned long a_entry;
-};
-
-#define N_TXTADDR(x)	0
-#define N_DATADDR(x)	0
-#define N_BSSADDR(x)	0
-#define N_DRSIZE(x)	0
-#define N_TRSIZE(x)	0
-#define N_SYMSIZE(x)	0
-#define N_TXTOFF(x)	0
-
-#endif /* _ASM_IA64_A_OUT_H */
diff --git a/include/asm-m32r/a.out.h b/include/asm-m32r/a.out.h
deleted file mode 100644
index ab150f5..0000000
--- a/include/asm-m32r/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef _ASM_M32R_A_OUT_H
-#define _ASM_M32R_A_OUT_H
-
-struct exec
-{
-  unsigned long a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* _ASM_M32R_A_OUT_H */
diff --git a/include/asm-m68knommu/a.out.h b/include/asm-m68knommu/a.out.h
deleted file mode 100644
index ce18ef9..0000000
--- a/include/asm-m68knommu/a.out.h
+++ /dev/null
@@ -1 +0,0 @@
-#include <asm-m68k/a.out.h>
diff --git a/include/asm-mips/a.out.h b/include/asm-mips/a.out.h
deleted file mode 100644
index cad8371..0000000
--- a/include/asm-mips/a.out.h
+++ /dev/null
@@ -1,35 +0,0 @@
-/*
- * This file is subject to the terms and conditions of the GNU General Public
- * License.  See the file "COPYING" in the main directory of this archive
- * for more details.
- *
- * Copyright (C) 1994 - 1999, 2003 by Ralf Baechle
- */
-#ifndef _ASM_A_OUT_H
-#define _ASM_A_OUT_H
-
-#ifdef __KERNEL__
-
-
-#endif
-
-struct exec
-{
-	unsigned long a_info;	/* Use macros N_MAGIC, etc for access */
-	unsigned a_text;	/* length of text, in bytes */
-	unsigned a_data;	/* length of data, in bytes */
-	unsigned a_bss;		/* length of uninitialized data area for
-				    file, in bytes */
-	unsigned a_syms;	/* length of symbol table data in file,
-				   in bytes */
-	unsigned a_entry;	/* start address */
-	unsigned a_trsize;	/* length of relocation info for text, in
-				    bytes */
-	unsigned a_drsize;	/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* _ASM_A_OUT_H */
diff --git a/include/asm-parisc/a.out.h b/include/asm-parisc/a.out.h
deleted file mode 100644
index eb04e34..0000000
--- a/include/asm-parisc/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef __PARISC_A_OUT_H__
-#define __PARISC_A_OUT_H__
-
-struct exec
-{
-  unsigned int a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* __A_OUT_GNU_H__ */
diff --git a/include/asm-powerpc/a.out.h b/include/asm-powerpc/a.out.h
deleted file mode 100644
index 89cead6..0000000
--- a/include/asm-powerpc/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef _ASM_POWERPC_A_OUT_H
-#define _ASM_POWERPC_A_OUT_H
-
-struct exec
-{
-	unsigned long a_info;	/* Use macros N_MAGIC, etc for access */
-	unsigned a_text;	/* length of text, in bytes */
-	unsigned a_data;	/* length of data, in bytes */
-	unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-	unsigned a_syms;	/* length of symbol table data in file, in bytes */
-	unsigned a_entry;	/* start address */
-	unsigned a_trsize;	/* length of relocation info for text, in bytes */
-	unsigned a_drsize;	/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* _ASM_POWERPC_A_OUT_H */
diff --git a/include/asm-sh/a.out.h b/include/asm-sh/a.out.h
deleted file mode 100644
index 1f93130..0000000
--- a/include/asm-sh/a.out.h
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef __ASM_SH_A_OUT_H
-#define __ASM_SH_A_OUT_H
-
-struct exec
-{
-  unsigned long a_info;		/* Use macros N_MAGIC, etc for access */
-  unsigned a_text;		/* length of text, in bytes */
-  unsigned a_data;		/* length of data, in bytes */
-  unsigned a_bss;		/* length of uninitialized data area for file, in bytes */
-  unsigned a_syms;		/* length of symbol table data in file, in bytes */
-  unsigned a_entry;		/* start address */
-  unsigned a_trsize;		/* length of relocation info for text, in bytes */
-  unsigned a_drsize;		/* length of relocation info for data, in bytes */
-};
-
-#define N_TRSIZE(a)	((a).a_trsize)
-#define N_DRSIZE(a)	((a).a_drsize)
-#define N_SYMSIZE(a)	((a).a_syms)
-
-#endif /* __ASM_SH_A_OUT_H */
diff --git a/include/asm-xtensa/a.out.h b/include/asm-xtensa/a.out.h
deleted file mode 100644
index fdf1370..0000000
--- a/include/asm-xtensa/a.out.h
+++ /dev/null
@@ -1,29 +0,0 @@
-/*
- * include/asm-xtensa/a.out.h
- *
- * Dummy a.out file. Xtensa does not support the a.out format, but the kernel
- * seems to depend on it.
- *
- * This file is subject to the terms and conditions of the GNU General Public
- * License.  See the file "COPYING" in the main directory of this archive
- * for more details.
- *
- * Copyright (C) 2001 - 2005 Tensilica Inc.
- */
-
-#ifndef _XTENSA_A_OUT_H
-#define _XTENSA_A_OUT_H
-
-struct exec
-{
-  unsigned long a_info;
-  unsigned a_text;
-  unsigned a_data;
-  unsigned a_bss;
-  unsigned a_syms;
-  unsigned a_entry;
-  unsigned a_trsize;
-  unsigned a_drsize;
-};
-
-#endif /* _XTENSA_A_OUT_H */

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Matthew Wilcox]

On Tue, Jun 17, 2008 at 11:42:35AM +0300, Adrian Bunk wrote:
> On Mon, Jun 16, 2008 at 01:26:54PM +0100, David Woodhouse wrote:
> > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > sense?
> 
> My guess would be that binfmt_aout isn't actually used there.

It was probably copied from MIPS in the early days of the parisc port
and nobody ever b othered to look at it before.  There was never an
a.out format for parisc.

-- 
Intel are signing my paycheques ... these opinions are still mine
"Bill, look, we understand that you're interested in selling us this
operating system, but compare it to ours.  We can't possibly take such
a retrograde step."

Re: architectures with ARCH_SUPPORTS_AOUT but no binfmt_aout

[Paul Mundt]

On Tue, Jun 17, 2008 at 11:42:35AM +0300, Adrian Bunk wrote:
> On Mon, Jun 16, 2008 at 01:26:54PM +0100, David Woodhouse wrote:
> >...
> > Actually, I think we can drop the preceding patch too -- it's not
> > necessary to fix the regression in 2.6.26, and dhowells suggests that we
> > might still want to use ARCH_SUPPORTS_AOUT, by making BINFMT_AOUT depend
> > on it instead of that hard-coded list of architectures.
> > 
> > If we do that, then ARCH_SUPPORTS_AOUT would need to be defined on a
> > different set of architectures to the ones it's currently defined on --
> > so I don't really want to go there before 2.6.26. I'll push just the
> > parts which were included in the mail I just sent.
> > 
> > Why _are_ there architectures which define ARCH_SUPPORTS_AOUT and have
> > <asm/a.out.h> but don't support binfmt_aout, anyway? How does that make
> > sense?
> 
> My guess would be that binfmt_aout isn't actually used there.
> 
> The affected architectures are:
> - h8300
> - m32r
> - parisc
> - sh
> - v850
> 
> v850 is broken beyond repair, and I hope my patch to remove the 
> architecture will make it into 2.6.27.
> 
> For the other 4 architectures I have Cc'ed the architecture maintainers.
> 
There was never an a.out format for any of sh/h8300/m32r. It's likely
just perpetuated damage copied over from the other ports, as in the
parisc case.