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
| ||
|
Date: Thu, 23 May 2019 13:32:54 +0800 From: Peter Xu <peterx@...hat.com> To: Frederic Weisbecker <frederic@...nel.org> Cc: linux-kernel@...r.kernel.org, Frederic Weisbecker <fweisbec@...il.com>, Thomas Gleixner <tglx@...utronix.de>, Ingo Molnar <mingo@...nel.org> Subject: Re: [PATCH] tick/sched: Drop duplicated tick_sched.inidle On Wed, May 22, 2019 at 02:18:41PM +0200, Frederic Weisbecker wrote: > On Wed, May 22, 2019 at 11:29:06AM +0800, Peter Xu wrote: > > It is set before entering idle and cleared when quitting idle, though > > it seems to be a complete duplicate of tick_sched.idle_active. We > > should probably be able to use any one of them to replace the other. > > Not exactly. > > @inidle is set on idle entry and cleared on idle exit. > @idle_active is the same but it's cleared during idle interrupts > so that idle_sleeptime only account real idle time. > > And note below: > > > @@ -1017,7 +1015,7 @@ void tick_nohz_irq_exit(void) > > { > > struct tick_sched *ts = this_cpu_ptr(&tick_cpu_sched); > > > > - if (ts->inidle) > > + if (ts->idle_active) > > tick_nohz_start_idle(ts); > > idle_active will always be cleared here from tick_nohz_irq_enter(). > We actually want to conditionally set it again depending on the inidle value. You are right; I've missed the calls from irq enter/exit. Thanks, Frederic. -- Peter Xu
Powered by blists - more mailing lists