[PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[Peter Huewe]

This patch fixes a build failure[1] for the avr32 architecture which seems
to be introduced by commit d7627467b7a8
"Make do_execve() take a const filename pointer"

The commit changes the definition in arch/avr32/kernel/process.c but
forgot to change it here.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
---
This build failure exists in linus' tree - added stable@kernel.org to the cc list

KernelVersion: linux-next-20101208
References:
[1] http://kisskb.ellerman.id.au/kisskb/buildresult/3607003/

 arch/avr32/include/asm/syscalls.h |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/avr32/include/asm/syscalls.h b/arch/avr32/include/asm/syscalls.h
index ab608b7..1440368 100644
--- a/arch/avr32/include/asm/syscalls.h
+++ b/arch/avr32/include/asm/syscalls.h
@@ -21,8 +21,10 @@ asmlinkage int sys_clone(unsigned long, unsigned long,
 			 unsigned long, unsigned long,
 			 struct pt_regs *);
 asmlinkage int sys_vfork(struct pt_regs *);
-asmlinkage int sys_execve(const char __user *, char __user *__user *,
-			  char __user *__user *, struct pt_regs *);
+asmlinkage int sys_execve(const char __user *ufilename,
+			  const char __user *const __user *uargv,
+			  const char __user *const __user *uenvp,
+			  struct pt_regs *regs);
 
 /* kernel/signal.c */
 asmlinkage int sys_sigaltstack(const stack_t __user *, stack_t __user *,
-- 
1.7.2.2

Re: [PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[Hans-Christian Egtvedt]

On Thu, 2010-12-09 at 00:19 +0100, Peter Huewe wrote: 
> This patch fixes a build failure[1] for the avr32 architecture which seems
> to be introduced by commit d7627467b7a8
> "Make do_execve() take a const filename pointer"
> 
> The commit changes the definition in arch/avr32/kernel/process.c but
> forgot to change it here.
> 
> Signed-off-by: Peter Huewe <peterhuewe@gmx.de>

Thanks for the heads-up.

> ---
> This build failure exists in linus' tree - added stable@kernel.org to the cc list
> 
> KernelVersion: linux-next-20101208
> References:
> [1] http://kisskb.ellerman.id.au/kisskb/buildresult/3607003/
> 
>  arch/avr32/include/asm/syscalls.h |    6 ++++--
>  1 files changed, 4 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/avr32/include/asm/syscalls.h b/arch/avr32/include/asm/syscalls.h
> index ab608b7..1440368 100644
> --- a/arch/avr32/include/asm/syscalls.h
> +++ b/arch/avr32/include/asm/syscalls.h
> @@ -21,8 +21,10 @@ asmlinkage int sys_clone(unsigned long, unsigned long,
>  			 unsigned long, unsigned long,
>  			 struct pt_regs *);
>  asmlinkage int sys_vfork(struct pt_regs *);
> -asmlinkage int sys_execve(const char __user *, char __user *__user *,
> -			  char __user *__user *, struct pt_regs *);
> +asmlinkage int sys_execve(const char __user *ufilename,
> +			  const char __user *const __user *uargv,
> +			  const char __user *const __user *uenvp,
> +			  struct pt_regs *regs);

Actually, I think it would be just as good to move to using the
syscalls.h in asm-generic. Let me prepare a patch or two for that. I can
try to push it through my, now hopefully working, arch tree.

-- 
Hans-Christian Egtvedt

Re: [stable] [PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[Greg KH]

On Thu, Dec 09, 2010 at 12:19:33AM +0100, Peter Huewe wrote:
> This patch fixes a build failure[1] for the avr32 architecture which seems
> to be introduced by commit d7627467b7a8
> "Make do_execve() take a const filename pointer"
> 
> The commit changes the definition in arch/avr32/kernel/process.c but
> forgot to change it here.
> 
> Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
> ---
> This build failure exists in linus' tree - added stable@kernel.org to the cc list

This is _not_ how to get stuff into the stable kernel tree, sorry.

Please read Documentation/stable_kernel_rules.txt for how to do so.

Hint, put:
	Cc: stable <stable@kernel.org.

in the signed-off-by: area of your patch and then you will automatically
get the patch applied to the stable trees when it goes to Linus.

Otherwise I have to dig through commit logs and hope to find it, which
wastes lots of time when it's not there yet.

Care to resend?

thanks,

greg k-h

Re: [stable] [PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[Peter Hüwe]

> This is _not_ how to get stuff into the stable kernel tree, sorry.
Sorry about that - still learning ;) - thanks for your patience.

> Please read Documentation/stable_kernel_rules.txt for how to do so.
 It states: "You must note the upstream commit ID in the changelog of your 
submission." - where do I get the ID from?

> Care to resend?
Will do.
Or should I rather drop the patch and we wait on Hans-Christian's patch(es)?

Thanks,
Peter

Re: [stable] [PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[Greg KH]

On Thu, Dec 09, 2010 at 07:42:36PM +0100, Peter Hüwe wrote:
> > This is _not_ how to get stuff into the stable kernel tree, sorry.
> Sorry about that - still learning ;) - thanks for your patience.
> 
> > Please read Documentation/stable_kernel_rules.txt for how to do so.
>  It states: "You must note the upstream commit ID in the changelog of your 
> submission." - where do I get the ID from?

That would be if the patch was in Linus's tree already.  Read the next
paragraph in the document:

 - To have the patch automatically included in the stable tree, add the tag
     Cc: stable@kernel.org
   in the sign-off area. Once the patch is merged it will be applied to
   the stable tree without anything else needing to be done by the author
   or subsystem maintainer.

> > Care to resend?
> Will do.
> Or should I rather drop the patch and we wait on Hans-Christian's patch(es)?

Whatever will end up in Linus's tree is the correct thing from a stable
point of view.

thanks,

greg k-h

[PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Peter Huewe]

This patch fixes a build failure[1] for the avr32 architecture which seems
to be introduced by commit d7627467b7a8
"Make do_execve() take a const filename pointer"

The commit changes the definition in arch/avr32/kernel/process.c but
forgot to change it here.

Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Cc: stable <stable@kernel.org>
---
KernelVersion: 20101209

 arch/avr32/include/asm/syscalls.h |    6 ++++--
 1 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/arch/avr32/include/asm/syscalls.h b/arch/avr32/include/asm/syscalls.h
index ab608b7..1440368 100644
--- a/arch/avr32/include/asm/syscalls.h
+++ b/arch/avr32/include/asm/syscalls.h
@@ -21,8 +21,10 @@ asmlinkage int sys_clone(unsigned long, unsigned long,
 			 unsigned long, unsigned long,
 			 struct pt_regs *);
 asmlinkage int sys_vfork(struct pt_regs *);
-asmlinkage int sys_execve(const char __user *, char __user *__user *,
-			  char __user *__user *, struct pt_regs *);
+asmlinkage int sys_execve(const char __user *ufilename,
+			  const char __user *const __user *uargv,
+			  const char __user *const __user *uenvp,
+			  struct pt_regs *regs);
 
 /* kernel/signal.c */
 asmlinkage int sys_sigaltstack(const stack_t __user *, stack_t __user *,
-- 
1.7.2.2

Re: [PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Hans-Christian Egtvedt]

On Thu, 2010-12-09 at 22:15 +0100, Peter Huewe wrote: 
> This patch fixes a build failure[1] for the avr32 architecture which seems
> to be introduced by commit d7627467b7a8
> "Make do_execve() take a const filename pointer"
> 
> The commit changes the definition in arch/avr32/kernel/process.c but
> forgot to change it here.
> 
> Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
> Cc: stable <stable@kernel.org>

Nack, I've piped up a patch in
http://git.kernel.org/?p=linux/kernel/git/egtvedt/avr32-2.6.git;a=shortlog;h=refs/heads/next that does this and more. The syscalls.h header file needs a bit more cleanup.

I'll also notify stable to get it backported. Do you have an idea how
far back it should go?

-- 
Hans-Christian Egtvedt

Re: [PATCH] avr32: Fix build failure conflicting types for 'sys_execve'

[David Howells]

Peter Huewe <peterhuewe@gmx.de> wrote:

> This patch fixes a build failure[1] for the avr32 architecture which seems
> to be introduced by commit d7627467b7a8
> "Make do_execve() take a const filename pointer"
> 
> The commit changes the definition in arch/avr32/kernel/process.c but
> forgot to change it here.
> 
> Signed-off-by: Peter Huewe <peterhuewe@gmx.de>

Acked-by: David Howells <dhowells@redhat.com>

Re: [PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Peter Hüwe]

Am Freitag 10 Dezember 2010, 11:25:41 schrieb Hans-Christian Egtvedt:
> Nack, I've piped up a patch in
> http://git.kernel.org/?p=linux/kernel/git/egtvedt/avr32-2.6.git;a=shortlog;
> h=refs/heads/next that does this and more. The syscalls.h header file needs
> a bit more cleanup.
> 
> I'll also notify stable to get it backported. Do you have an idea how
> far back it should go?

Hi,

Linus' tree fails since August 19th due to this issue, so it falls into the 
2.6.36 merge window only.
http://kisskb.ellerman.id.au/kisskb/buildresult/3059836/
(but it still failed for quite a while longer due to other issues ;)


Thanks,
Peter
(a reported by on the patch would have been nice ;)

Re: [PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Hans-Christian Egtvedt]

On Sun, 2010-12-12 at 20:37 +0100, Peter Hüwe wrote: 
> Am Freitag 10 Dezember 2010, 11:25:41 schrieb Hans-Christian Egtvedt:
> > Nack, I've piped up a patch in
> > http://git.kernel.org/?p=linux/kernel/git/egtvedt/avr32-2.6.git;a=shortlog;
> > h=refs/heads/next that does this and more. The syscalls.h header file needs
> > a bit more cleanup.
> > 
> > I'll also notify stable to get it backported. Do you have an idea how
> > far back it should go?
> 
> Hi,
> 
> Linus' tree fails since August 19th due to this issue, so it falls into the 
> 2.6.36 merge window only.
> http://kisskb.ellerman.id.au/kisskb/buildresult/3059836/

Thanks for digging that up. The patch will only need to be applied to
v2.6.36 and beyond then.

> (but it still failed for quite a while longer due to other issues ;)

What other issues?

> Thanks,
> Peter
> (a reported by on the patch would have been nice ;)

Of course, I'll rebase and add you're name to it. Does the same go for
David Howells?

-- 
Hans-Christian Egtvedt

Re: [PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Peter Hüwe]

Am Montag 13 Dezember 2010, 09:12:53 schrieb Hans-Christian Egtvedt:
> > (but it still failed for quite a while longer due to other issues ;)
> What other issues?

There's a linker error at the end:

  CC      init/version.o
  LD      init/built-in.o
  LD      vmlinux
lib/lib.a(vsprintf.o): In function `resource_string':
vsprintf.c:(.text+0x149e): relocation truncated to fit: R_AVR32_9UW_PCREL 
against `.text'+16a8
vsprintf.c:(.text+0x14ae): relocation truncated to fit: R_AVR32_9UW_PCREL 
against `.text'+16ac

which seems to exist since May 26th in Linus' tree.
See 
http://kisskb.ellerman.id.au/kisskb/buildresult/2631302/
http://kisskb.ellerman.id.au/kisskb/target/664/
for more details.

Unfortunately I wasn't able to locate the offending commit using git bisect :/



Reported-by: Peter Huewe <peterhuewe@gmx.de>

> > (a reported by on the patch would have been nice ;)
> Of course, I'll rebase and add you're name to it. Does the same go for
> David Howells?
Great, thanks. - although I'm not sure wether an ack, reported-by or signed-
off by fits best ;)
I did get an ack from him:
> Acked-by: David Howells <dhowells@redhat.com>


Thanks,
Peter

Re: [PATCH v2] avr32: Fix build failure conflicting types for 'sys_execve'

[Hans-Christian Egtvedt]

On Mon, 2010-12-13 at 18:12 +0100, Peter Hüwe wrote: 
> Am Montag 13 Dezember 2010, 09:12:53 schrieb Hans-Christian Egtvedt:
> > > (but it still failed for quite a while longer due to other issues ;)
> > What other issues?
> 
> There's a linker error at the end:
> 
>   CC      init/version.o
>   LD      init/built-in.o
>   LD      vmlinux
> lib/lib.a(vsprintf.o): In function `resource_string':
> vsprintf.c:(.text+0x149e): relocation truncated to fit: R_AVR32_9UW_PCREL 
> against `.text'+16a8
> vsprintf.c:(.text+0x14ae): relocation truncated to fit: R_AVR32_9UW_PCREL 
> against `.text'+16ac
> 
> which seems to exist since May 26th in Linus' tree.
> See 
> http://kisskb.ellerman.id.au/kisskb/buildresult/2631302/
> http://kisskb.ellerman.id.au/kisskb/target/664/
> for more details.
> 
> Unfortunately I wasn't able to locate the offending commit using git bisect :/

Hmm, I haven't seen that, which toolchain and which configuration do you
use?

-- 
Hans-Christian Egtvedt