Messages in this thread | | | Subject | Re: [PATCH v2 7/9] sched,debug: Convert sysctl sched_domains to debugfs | From | Dietmar Eggemann <> | Date | Thu, 15 Apr 2021 14:16:14 +0200 |
| |
On 15/04/2021 11:06, Peter Zijlstra wrote: > On Tue, Apr 13, 2021 at 03:55:15PM +0100, Valentin Schneider wrote: >> On 12/04/21 12:14, Peter Zijlstra wrote: >>> Stop polluting sysctl, move to debugfs for SCHED_DEBUG stuff. >>> >>> Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org> >>> Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com> >> >> On my Juno (2+4 big.LITTLE), sys/kernel/debug/sched/domains/ is now empty. >> >> I think that's because of unregister_sched_domain_sysctl() - >> debugfs_remove() is recursive, and I do get a case where we rebuild the >> domains but no CPU has been added or removed (we rebuild the domains when >> cpufreq kicks in, it's part of the big.LITTLE ponies). >> >> Do we actually still need that unregister? From a brief glance it looks >> like we could throw it out. > > Yeah, I can't think of anything either. AFAICT it hasn't done anything > useful since that cpumask optimization. Consider it gone. > > I'll let it soak for another day or so, but then I was planning on > merging this series. > > Updated patch has been in queue.git/sched/debug since yesterday.
Had to check since v1 was working fine on Juno. So it was this __cpumask_clear_cpu() in register_sched_domain_sysctl() introduced in v2 which let the files under /sys/kernel/debug/sched/domains disapear.
With {un,}register_sched_domain_sysctl() removed from partition_sched_domains_locked() they're there again.
Looks like now register_sched_domain_sysctl() can be made static in kernel/sched/debug.c.
[...]
| |