Re: [PATCH] save/restore errno around pthread_atfork handlers

[Rich Felker <dalias@libc.org>]

On Fri, Nov 10, 2017 at 02:58:29PM -0600, Bobby Bingham wrote:
> If the syscall fails, errno must be preserved for the caller. There's no
> guarantee that the handlers registered with pthread_atfork won't clobber
> errno.
> ---
>  src/process/fork.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/src/process/fork.c b/src/process/fork.c
> index b96f0024..6602eafc 100644
> --- a/src/process/fork.c
> +++ b/src/process/fork.c
> @@ -15,6 +15,7 @@ pid_t fork(void)
>  {
>  	pid_t ret;
>  	sigset_t set;
> +	int olderr;
>  	__fork_handler(-1);
>  	__block_all_sigs(&set);
>  #ifdef SYS_fork
> @@ -30,6 +31,10 @@ pid_t fork(void)
>  		libc.threads_minus_1 = 0;
>  	}
>  	__restore_sigs(&set);
> +
> +	olderr = errno;
>  	__fork_handler(!ret);
> +	errno = olderr;
> +
>  	return ret;
>  }
> -- 
> 2.15.0

I think the patch as written is incorrect, because it can set errno to
0 after application code in the atfork handler set it to something
nonzero; doing so is non-conforming.

It would be possible to special-case to avoid this, but it probably
makes more sense to just call SYS_fork/SYS_clone with __syscall rather
than syscall, then return __syscall_ret(ret) instead of return ret.
Does that sound correct?

Rich