Re: [PATCH v2] sched/deadline: Remove fair-servers from real-time task's bandwidth accounting

[Matteo Martelli]

Hi Yuri,

On Fri, 25 Jul 2025 18:44:12 +0200, Yuri Andriaccio <yurand2000@xxxxxxxxx> wrote:
> ...
> @@ -1688,17 +1690,14 @@ int dl_server_apply_params(struct sched_dl_entity *dl_se, u64 runtime, u64 perio
>
>	cpus = dl_bw_cpus(cpu);
>	cap = dl_bw_capacity(cpu);
> +	max_bw = div64_ul(cap_scale(BW_UNIT - dl_b->bw, cap), (unsigned long)cpus);

This line exceeds 80 characters width. Perhaps it needs to be split.

> 
> -	if (__dl_overflow(dl_b, cap, old_bw, new_bw))
> +	if (new_bw > max_bw)
> 		return -EBUSY;
> ...

Beside that minor note, I retested the v2 of this patch with the same
tests I ran for v1 [1]. I confirm that stress-ng and runtime variations
commands provide the same results. Also no warning is produced anymore
as I also applied Juri's patch as you suggested [2].

Tested-by: Matteo Martelli <matteo.martelli@xxxxxxxxxxxxxxx>

[1]: https://lore.kernel.org/all/86013fcc38e582ab89b9b7e4864cc1bd@xxxxxxxxxxxxxxx/
[2]: https://lore.kernel.org/all/20250725152804.14224-1-yurand2000@xxxxxxxxx/