Re: [PATCH] reintroduce accept4

[Andrew Morton <akpm@linux-foundation.org>]

(cc linux-api)

(cc linux-arch)

On Sun, 26 Oct 2008 12:41:50 -0400 Ulrich Drepper <drepper@redhat.com> wrote:

> This patch reintroduces accept4, replacing paccept.  It's easy to see that
> the patch only removes code and then redirects existing code away from the
> removed functions.  Since the paccept code sans signal handling was never
> in question I think there is no reason to quarantine the patch first.

I'll confess to not having a clue what's going on here.

What is accept4() and why do I want one?  Sigh.  Hopefully others have
been following more closely and have some context.

> I've updated the test program which now looks as follows:

> 
> #include <fcntl.h>
> #include <pthread.h>
> #include <stdio.h>
> #include <unistd.h>
> #include <netinet/in.h>
> #include <sys/socket.h>
> #include <sys/syscall.h>
> 
> #ifdef __x86_64__
> #define __NR_accept4 288
> #define SOCK_CLOEXEC O_CLOEXEC
> #elif __i386__
> #define SYS_ACCEPT4     18
> #define USE_SOCKETCALL 1
> #define SOCK_CLOEXEC O_CLOEXEC
> #else

Well.  This doesn't actually agree with the kernel patch.

>
> ...
>
>  arch/x86/include/asm/unistd_64.h |    4 -
>  include/linux/net.h              |    6 --
>  include/linux/syscalls.h         |    3 -
>  kernel/sys_ni.c                  |    2 
>  net/compat.c                     |   50 ++----------------------
>  net/socket.c                     |   80 ++++-----------------------------------

I'd suggest that i386 is sufficiently common to warrant its inclusion
in the initial patch.