Re: xtensa: switch to generic clone()

Re: xtensa: switch to generic clone()

[Max Filippov]

Hi Al,

looks like the following hunk is missing in the commit
  a4bcde6 xtensa: switch to generic clone()
found in the linux-next tree next-20121126, which breaks xtensa builds.

---
diff --git a/arch/xtensa/include/asm/syscall.h
b/arch/xtensa/include/asm/syscall.h
index b00c928..cd53ec2 100644
--- a/arch/xtensa/include/asm/syscall.h
+++ b/arch/xtensa/include/asm/syscall.h
@@ -8,6 +8,8 @@
  * Copyright (C) 2001 - 2007 Tensilica Inc.
  */

+#include <asm-generic/syscalls.h>
+
 struct pt_regs;
 struct sigaction;
 asmlinkage long xtensa_ptrace(long, long, long, long);


-- 
Thanks.
-- Max

Re: xtensa: switch to generic clone()

[Al Viro]

On Tue, Nov 27, 2012 at 08:48:24AM +0300, Max Filippov wrote:
> Hi Al,
> 
> looks like the following hunk is missing in the commit
>   a4bcde6 xtensa: switch to generic clone()
> found in the linux-next tree next-20121126, which breaks xtensa builds.
> 
> ---
> diff --git a/arch/xtensa/include/asm/syscall.h
> b/arch/xtensa/include/asm/syscall.h
> index b00c928..cd53ec2 100644
> --- a/arch/xtensa/include/asm/syscall.h
> +++ b/arch/xtensa/include/asm/syscall.h
> @@ -8,6 +8,8 @@
>   * Copyright (C) 2001 - 2007 Tensilica Inc.
>   */
> 
> +#include <asm-generic/syscalls.h>

The bug is real, all right, but I'm not sure I like that fix.
The problem hits in arch/xtensa/kernel/syscall.c, right?  I'd rather
get correct declaration into include/linux/syscalls.h and drop the
asm-generic/syscalls.h one.  I'll do it that way the next time I
push into for-next (in the morning, that is); for now your variant
will do as a temporary fix, of course.

Incidentally, xtensa_shmat() looks very similar to sys_shmat().  The
only difference is force_successful_syscall_return() in the latter
and that expands to do { } while (0) on xtensa (i.e. default gets used).
Granted, I'm half asleep right now and could've easily missed something,
but...