Re: [PATCH v4] signal: trace_signal_deliver when signal_group_exit

[Christian Brauner]

On Wed, Apr 24, 2019 at 08:52:38PM +0800, Zhenliang Wei wrote:
> In the fixes commit, removing SIGKILL from each thread signal mask
> and executing "goto fatal" directly will skip the call to
> "trace_signal_deliver". At this point, the delivery tracking of the SIGKILL
> signal will be inaccurate.
> 
> Therefore, we need to add trace_signal_deliver before "goto fatal"
> after executing sigdelset.
> 
> Note: SEND_SIG_NOINFO matches the fact that SIGKILL doesn't have any info.
> 
> Acked-by: Christian Brauner <christian@xxxxxxxxxx>

I think we're supposed to use more Reviewed-bys so feel free (or Andrew)
to change this to:

Reviewed-by: Christian Brauner <christian@xxxxxxxxxx>

> Reviewed-by: Oleg Nesterov <oleg@xxxxxxxxxx>
> Cc: <stable@xxxxxxxxxxxxxxx>
> Fixes: cf43a757fd4944 ("signal: Restore the stop PTRACE_EVENT_EXIT")
> Reported-by: kbuild test robot <lkp@xxxxxxxxx>
> Signed-off-by: Zhenliang Wei <weizhenliang@xxxxxxxxxx>
> ---
>  kernel/signal.c | 2 ++
>  1 file changed, 2 insertions(+)
> 
> diff --git a/kernel/signal.c b/kernel/signal.c
> index 227ba170298e..3edf526db7c6 100644
> --- a/kernel/signal.c
> +++ b/kernel/signal.c
> @@ -2441,6 +2441,8 @@ bool get_signal(struct ksignal *ksig)
>  	if (signal_group_exit(signal)) {
>  		ksig->info.si_signo = signr = SIGKILL;
>  		sigdelset(&current->pending.signal, SIGKILL);
> +		trace_signal_deliver(SIGKILL, SEND_SIG_NOINFO,
> +			&sighand->action[signr - 1]);

Hm, sorry for being the really nitpicky person here. Just for the sake
of consistency how about we do either:

+		trace_signal_deliver(SIGKILL, SEND_SIG_NOINFO,
+			&sighand->action[SIGKILL - 1]);

or

+		trace_signal_deliver(signr, SEND_SIG_NOINFO,
+			&sighand->action[signr - 1]);

I'm not going to argue about this though. Can just also leave it as is.

Christian