[PATCH 1/4] Consolidate Linux accept implementation

[Phil Blundell]

On Mon, 2017-01-30 at 12:02 -0200, Adhemerval Zanella wrote:
> 
> Â  2. Define __NR_acccept as default (__ASSUME_ACCEPT_SYSCALL) and
> undef for

Too many "c"s in accept?

> diff --git a/sysdeps/unix/sysv/linux/m68k/kernel-features.h
> b/sysdeps/unix/sysv/linux/m68k/kernel-features.h
> index 6811a28..f583ae7 100644
> --- a/sysdeps/unix/sysv/linux/m68k/kernel-features.h
> +++ b/sysdeps/unix/sysv/linux/m68k/kernel-features.h
> @@ -27,7 +27,6 @@
> Â # define __ASSUME_BIND_SYSCALLÂ Â Â Â Â Â Â Â Â Â Â Â Â Â Â 1
> Â # define __ASSUME_CONNECT_SYSCALLÂ Â Â Â Â Â Â Â Â Â Â Â 1
> Â # define __ASSUME_LISTEN_SYSCALLÂ Â Â Â Â Â Â Â Â Â Â Â Â 1
> -# define __ASSUME_ACCEPT4_FOR_ACCEPT_SYSCALL 1
> Â # define __ASSUME_GETSOCKOPT_SYSCALLÂ Â Â Â Â Â Â Â Â 1
> Â # define __ASSUME_SETSOCKOPT_SYSCALLÂ Â Â Â Â Â Â Â Â 1
> Â # define __ASSUME_GETSOCKNAME_SYSCALLÂ Â Â Â Â Â Â Â 1
> @@ -41,7 +40,10 @@
> Â 
> Â #include_next <kernel-features.h>
> Â 
> +#undef __ASSUME_ACCEPT_SYSCALL
> +
> Â #if __LINUX_KERNEL_VERSION < 0x040300
> +# undef __ASSUME_ACCCEPT_SYSCALL

This is a bit odd. Â __ASSUME_ACCCEPT_SYSCALL never seems to be defined
anywhere that I can find, but you are explicitly undefining it for
these old kernels. Â I would have assumed it was a typo for
__ASSUME_ACCEPT_SYSCALL if you weren't also undefining that one a
couple of lines further up. Â Can you clarify?

p.