Re: [PATCH v2] exec: don't force_sigsegv processes with a pending fatal signal

[Oleg Nesterov <oleg@redhat.com>]

sorry, I couldn't look at this patch before.

On 02/04, Ivan Delalande wrote:
>
> --- a/fs/exec.c
> +++ b/fs/exec.c
> @@ -1660,7 +1660,12 @@ int search_binary_handler(struct linux_binprm *bprm)
>  		if (retval < 0 && !bprm->mm) {
>  			/* we got to flush_old_exec() and failed after it */
>  			read_unlock(&binfmt_lock);
> -			force_sigsegv(SIGSEGV, current);
> +			if (!fatal_signal_pending(current)) {
> +				if (print_fatal_signals)
> +					pr_info("load_binary() failed: %d\n",
> +						retval);

I won't argue, but do we really want this spam?

> +				force_sigsegv(SIGSEGV, current);
> +			}
>  			return retval;
>  		}
>  		if (retval != -ENOEXEC || !bprm->file) {
> diff --git a/kernel/signal.c b/kernel/signal.c
> index e1d7ad8e6ab1..674076e63624 100644
> --- a/kernel/signal.c
> +++ b/kernel/signal.c
> @@ -2552,10 +2552,10 @@ static void signal_delivered(struct ksignal *ksig, int stepping)
>
>  void signal_setup_done(int failed, struct ksignal *ksig, int stepping)
>  {
> -	if (failed)
> -		force_sigsegv(ksig->sig, current);
> -	else
> +	if (!failed)
>  		signal_delivered(ksig, stepping);
> +	else if (!fatal_signal_pending(current))
> +		force_sigsegv(ksig->sig, current);

The changelog doesn't explain this change.

OK, I guess it comes from the previous discussion, setup_rt_frame() can equally fail
if fatal_signal_pending(). But this should be documented at least in the changelog,
and I still think we could simply change force_sigsegv() instead.

In any case, Eric has already mentioned that we going to give SIGKILL more priority,
so I think we can drop this patch?

Oleg.