linux-kernel - Re: [PATCH 09/16] rcu: Remove redundant check for online cpu

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]
Message-ID: <20140723153018.GY11241@linux.vnet.ibm.com>
Date:	Wed, 23 Jul 2014 08:30:18 -0700
From:	"Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
To:	Pranith Kumar <bobby.prani@...il.com>
Cc:	Josh Triplett <josh@...htriplett.org>,
	Steven Rostedt <rostedt@...dmis.org>,
	Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
	Lai Jiangshan <laijs@...fujitsu.com>,
	"open list:READ-COPY UPDATE..." <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 09/16] rcu: Remove redundant check for online cpu

On Wed, Jul 23, 2014 at 11:11:45AM -0400, Pranith Kumar wrote:
> On Wed, Jul 23, 2014 at 10:23 AM, Paul E. McKenney
> <paulmck@...ux.vnet.ibm.com> wrote:
> > On Wed, Jul 23, 2014 at 10:12:54AM -0400, Pranith Kumar wrote:
> >> On Wed, Jul 23, 2014 at 9:50 AM, Paul E. McKenney
> >> <paulmck@...ux.vnet.ibm.com> wrote:
> >> > On Wed, Jul 23, 2014 at 08:59:06AM -0400, Pranith Kumar wrote:
> >> >> On Wed, Jul 23, 2014 at 8:21 AM, Paul E. McKenney
> >> >> <paulmck@...ux.vnet.ibm.com> wrote:
> >> >> > On Wed, Jul 23, 2014 at 01:09:46AM -0400, Pranith Kumar wrote:
> >> >> >> There are two checks for an online CPU if two if() conditions. This commit
> >> >> >> simplies this by replacing it with only one check for the online CPU.
> >> >> >>
> >> >> >> Signed-off-by: Pranith Kumar <bobby.prani@...il.com>
> >> >> >
> >> >> > I admit that it is very early in the morning my time, but I don't see
> >> >> > this change as preserving the semantics in all cases.  Please recheck
> >> >> > your changes to the second check.
> >> >> >
> >> >> >                                                         Thanx, Paul
> >> >>
> >> >> I guess you must be thrown off by the complementary checks, the first
> >> >> check is for cpu_online() and second is for cpu_is_offline(). :)
> >> >>
> >> >> Previously, if a cpu is offline, the first condition is false and the
> >> >> second condition is true, so we return from the second if() condition.
> >> >> The same semantics are being preserved.
> >> >
> >> > Fair enough!
> >> >
> >> > Nevertheless, I am not seeing this as a simplification.
> >>
> >> I am not sure what you mean here, do you mean that both the checks are
> >> actually required?
> >
> > I mean that the current compound tests each mean something.  Pulling out
> > the offline test adds lines of code and obscures that meaning.   This means
> > that it is easier (for me, anyway) to see why the current code is correct
> > than it is to see why your suggested change is correct.
> >
> 
> That is a valid point. I did not mean to reduce readability of the
> code. Just trying to avoid the overhead of smp_processor_id().
> 
> Not sure if you would prefer this, but how about the following?

If you change the "awake" to something like "am_online", I could get
behind this one.

							Thanx, Paul

> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> index f1ba773..3a26008 100644
> --- a/kernel/rcu/tree.c
> +++ b/kernel/rcu/tree.c
> @@ -2516,15 +2516,16 @@ static void __call_rcu_core(struct rcu_state
> *rsp, struct rcu_data *rdp,
>  {
>         bool needwake;
> 
> +       bool awake = cpu_online(smp_processor_id);
>         /*
>          * If called from an extended quiescent state, invoke the RCU
>          * core in order to force a re-evaluation of RCU's idleness.
>          */
> -       if (!rcu_is_watching() && cpu_online(smp_processor_id()))
> +       if (!rcu_is_watching() && awake)
>                 invoke_rcu_core();
> 
>         /* If interrupts were disabled or CPU offline, don't invoke RCU core. */
> -       if (irqs_disabled_flags(flags) || cpu_is_offline(smp_processor_id()))
> +       if (irqs_disabled_flags(flags) || !awake)
>                 return;
> 
>         /*
> 
> 
> >
> >> >> --
> >> >> Pranith.
> >> >>
> >> >>
> >> >>
> >> >> >
> >> >> >> ---
> >> >> >>  kernel/rcu/tree.c | 9 ++++++---
> >> >> >>  1 file changed, 6 insertions(+), 3 deletions(-)
> >> >> >>
> >> >> >> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
> >> >> >> index 5dcbf36..8d598a2 100644
> >> >> >> --- a/kernel/rcu/tree.c
> >> >> >> +++ b/kernel/rcu/tree.c
> >> >> >> @@ -2602,15 +2602,18 @@ static void __call_rcu_core(struct rcu_state *rsp, struct rcu_data *rdp,
> >> >> >>  {
> >> >> >>       bool needwake;
> >> >> >>
> >> >> >> +     if (!cpu_online(smp_processor_id()))
> >> >> >> +             return;
> >> >> >> +
> >> >> >>       /*
> >> >> >>        * If called from an extended quiescent state, invoke the RCU
> >> >> >>        * core in order to force a re-evaluation of RCU's idleness.
> >> >> >>        */
> >> >> >> -     if (!rcu_is_watching() && cpu_online(smp_processor_id()))
> >> >> >> +     if (!rcu_is_watching())
> >> >> >>               invoke_rcu_core();
> >> >> >>
> >> >> >> -     /* If interrupts were disabled or CPU offline, don't invoke RCU core. */
> >> >> >> -     if (irqs_disabled_flags(flags) || cpu_is_offline(smp_processor_id()))
> >> >> >> +     /* If interrupts were disabled, don't invoke RCU core. */
> >> >> >> +     if (irqs_disabled_flags(flags))
> >> >> >>               return;
> >> >> >>
> >> >> >>       /*
> >> >> >> --
> >> >> >> 2.0.0.rc2
> >> >> >>
> >> >> >
> >> >>
> >> >>
> >> >>
> >> >> --
> >> >> Pranith
> >> >>
> >> >
> >>
> >>
> >>
> >> --
> >> Pranith
> >>
> >
> 
> 
> 
> -- 
> Pranith
> 

--
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/