linux-kernel - Re: [PATCH v2 5/6] X86: remove redundant cpuidle_idle

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 for Android: free password hash cracker in your pocket

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

Message-ID: <alpine.LFD.2.11.1401291434120.1652@knanqh.ubzr>
Date:	Wed, 29 Jan 2014 15:14:40 -0500 (EST)
From:	Nicolas Pitre <nicolas.pitre@...aro.org>
To:	Olof Johansson <olof@...om.net>
cc:	Russell King <linux@....linux.org.uk>,
	Benjamin Herrenschmidt <benh@...nel.crashing.org>,
	Preeti U Murthy <preeti@...ux.vnet.ibm.com>,
	Paul Mundt <lethal@...ux-sh.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	Ingo Molnar <mingo@...hat.com>,
	Peter Zijlstra <peterz@...radead.org>,
	"Rafael J. Wysocki" <rjw@...ysocki.net>,
	Daniel Lezcano <daniel.lezcano@...aro.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"linux-arm-kernel@...ts.infradead.org" 
	<linux-arm-kernel@...ts.infradead.org>,
	linuxppc-dev <linuxppc-dev@...ts.ozlabs.org>,
	Linux-sh list <linux-sh@...r.kernel.org>,
	linux-pm@...r.kernel.org,
	"linaro-kernel@...ts.linaro.org" <linaro-kernel@...ts.linaro.org>
Subject: Re: [PATCH v2 5/6] X86: remove redundant cpuidle_idle_call()

On Wed, 29 Jan 2014, Olof Johansson wrote:

> Hi,
> 
> On Wed, Jan 29, 2014 at 9:45 AM, Nicolas Pitre <nicolas.pitre@...aro.org> wrote:
> > The core idle loop now takes care of it.
> >
> > Signed-off-by: Nicolas Pitre <nico@...aro.org>
> > Acked-by: Daniel Lezcano <daniel.lezcano@...aro.org>
> > ---
> >  arch/x86/kernel/process.c | 5 +----
> >  1 file changed, 1 insertion(+), 4 deletions(-)
> >
> > diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c
> > index 3fb8d95ab8..4505e2a950 100644
> > --- a/arch/x86/kernel/process.c
> > +++ b/arch/x86/kernel/process.c
> > @@ -298,10 +298,7 @@ void arch_cpu_idle_dead(void)
> >   */
> >  void arch_cpu_idle(void)
> >  {
> > -       if (cpuidle_idle_call())
> > -               x86_idle();
> > -       else
> > -               local_irq_enable();
> > +       x86_idle();
> 
> You're taking out the local_irq_enable() here but I don't see the
> equivalent of adding it back in the 1/6 patch that moves the
> cpuidle_idle_call() up to common code. It seems that one of the call
> paths through cpuidle_idle_call() don't re-enable it on its own.

When cpuidle_idle_call() returns non-zero, IRQs are left disabled.  When 
it returns zero then IRQs should be disabled.  Same goes for cpuidle 
drivers.  That's the theory at least.

Looking into some cpuidle drivers for x86 I found at least one that 
doesn't respect this convention.  Damn.

> Even if this is the right thing to do, why it's OK to do so should
> probably be documented in the patch description.

Better yet, I'm going to amend patch 1/6 with the below to make things 
more reliable while still identifying misbehaving drivers.

diff --git a/kernel/cpu/idle.c b/kernel/cpu/idle.c
index ffcd3ee9af..14ca43430a 100644
--- a/kernel/cpu/idle.c
+++ b/kernel/cpu/idle.c
@@ -98,7 +98,8 @@ static void cpu_idle_loop(void)
 					rcu_idle_enter();
 					if (cpuidle_idle_call())
 						arch_cpu_idle();
-					WARN_ON_ONCE(irqs_disabled());
+					if (WARN_ON_ONCE(irqs_disabled()))
+						local_irq_enable();
 					rcu_idle_exit();
 					start_critical_timings();
 				} else {
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/