linux-kernel - Re: [PATCH v4 03/10] sched/topology: Propagate SD_ASYM

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 6 Aug 2020 16:20:25 +0200
From:   Ingo Molnar <mingo@...nel.org>
To:     Valentin Schneider <valentin.schneider@....com>
Cc:     linux-kernel@...r.kernel.org, Quentin Perret <qperret@...gle.com>,
        peterz@...radead.org, vincent.guittot@...aro.org,
        dietmar.eggemann@....com, morten.rasmussen@....com
Subject: Re: [PATCH v4 03/10] sched/topology: Propagate SD_ASYM_CPUCAPACITY
 upwards


* Valentin Schneider <valentin.schneider@....com> wrote:

> We currently set this flag *only* on domains whose topology level exactly
> match the level where we detect asymmetry (as returned by
> asym_cpu_capacity_level()). This is rather problematic.
> 
> Say there are two clusters in the system, one with a lone big CPU and the
> other with a mix of big and LITTLE CPUs (as is allowed by DynamIQ):
> 
> DIE [                ]
> MC  [             ][ ]
>      0   1   2   3  4
>      L   L   B   B  B
> 
> asym_cpu_capacity_level() will figure out that the MC level is the one
> where all CPUs can see a CPU of max capacity, and we will thus set
> SD_ASYM_CPUCAPACITY at MC level for all CPUs.
> 
> That lone big CPU will degenerate its MC domain, since it would be alone in
> there, and will end up with just a DIE domain. Since the flag was only set
> at MC, this CPU ends up not seeing any SD with the flag set, which is
> broken.
> 
> Rather than clearing dflags at every topology level, clear it before
> entering the topology level loop. This will properly propagate upwards
> flags that are set starting from a certain level.
> 
> Reviewed-by: Quentin Perret <qperret@...gle.com>
> Signed-off-by: Valentin Schneider <valentin.schneider@....com>
> ---
>  kernel/sched/topology.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c
> index 865fff3ef20a..42b89668e1e4 100644
> --- a/kernel/sched/topology.c
> +++ b/kernel/sched/topology.c
> @@ -1985,11 +1985,10 @@ build_sched_domains(const struct cpumask *cpu_map, struct sched_domain_attr *att
>  	/* Set up domains for CPUs specified by the cpu_map: */
>  	for_each_cpu(i, cpu_map) {
>  		struct sched_domain_topology_level *tl;
> +		int dflags = 0;
>  
>  		sd = NULL;
>  		for_each_sd_topology(tl) {
> -			int dflags = 0;
> -
>  			if (tl == tl_asym) {
>  				dflags |= SD_ASYM_CPUCAPACITY;
>  				has_asym = true;

I'd suggest ordering all patches with potential side effects at the 
end, to make them easier to bisect.

I.e. I'd reorder this series to do:

 - Obviously correct renamings & cleanups

 - Convert the code over to the new instrumented sd-flags method. This 
   will presumably spew a few warnings for problems the new debugging 
   checks catch in existing topologies.

 - Do all the behavioral changes and fixes like this patch, even if we 
   think that they have no serious side effects.

In that sense it might make sense to order the two ARM patches to the 
later stage as well - but I suppose it's OK to do those two first as 
well.

Nice series otherwise, these new checks look really useful and already 
caught bugs.

Thanks,

	Ingo