Re: [patch v6 02/20] posix-timers: Make signal overrun accounting sensible

[Frederic Weisbecker <frederic@kernel.org>]

Le Thu, Oct 31, 2024 at 04:46:25PM +0100, Thomas Gleixner a écrit :
> The handling of the timer overrun in the signal code is inconsistent as it
> takes previous overruns into account. This is just wrong as after the
> reprogramming of a timer the overrun count starts over from a clean state,
> i.e. 0.
> 
> Don't touch info::si_overrun in send_sigqueue() and only store the overrun
> value at signal delivery time, which is computed from the timer itself
> relative to the expiry time.
> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> ---
> V6: Fold the timer_overrun_to_int() cleanup from Frederic and remove all
>     overrun fiddling from the signal path.
> ---
>  kernel/signal.c            |    6 ------
>  kernel/time/posix-timers.c |   11 ++++++-----
>  2 files changed, 6 insertions(+), 11 deletions(-)
> ---
> 
> --- a/kernel/signal.c
> +++ b/kernel/signal.c
> @@ -1968,15 +1968,9 @@ int send_sigqueue(struct sigqueue *q, st
>  
>  	ret = 0;
>  	if (unlikely(!list_empty(&q->list))) {
> -		/*
> -		 * If an SI_TIMER entry is already queue just increment
> -		 * the overrun count.
> -		 */
> -		q->info.si_overrun++;
>  		result = TRACE_SIGNAL_ALREADY_PENDING;
>  		goto out;
>  	}
> -	q->info.si_overrun = 0;

So it's not cleared anymore on signal queue?

Not sure if it's a big problem but if an interval timer gets a signal with
overruns and then the timer is reset later as non interval, the resulting
upcoming signals will still carry the previous non-zero overruns?

However it's better to keep the overrun update on a single place so
perhaps this?

diff --git a/kernel/time/posix-timers.c b/kernel/time/posix-timers.c
index 66ed49efc02f..f06c52731d65 100644
--- a/kernel/time/posix-timers.c
+++ b/kernel/time/posix-timers.c
@@ -282,6 +282,8 @@ bool posixtimer_deliver_signal(struct kernel_siginfo *info)
 		++timr->it_signal_seq;
 
 		info->si_overrun = timer_overrun_to_int(timr);
+	} else {
+		info->si_overrun = 0;
 	}
 	ret = true;
 
Other than that:

Reviewed-by: Frederic Weisbecker <frederic@kernel.org>