[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20110917190538.GE2333@linux.vnet.ibm.com>
Date: Sat, 17 Sep 2011 12:05:38 -0700
From: "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
To: Frederic Weisbecker <fweisbec@...il.com>
Cc: LKML <linux-kernel@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Peter Zijlstra <a.p.zijlstra@...llo.nl>,
"H. Peter Anvin" <hpa@...or.com>,
Andy Henroid <andrew.d.henroid@...el.com>,
Mike Frysinger <vapier@...too.org>,
Guan Xuetao <gxt@...c.pku.edu.cn>,
David Miller <davem@...emloft.net>,
Chris Metcalf <cmetcalf@...era.com>,
Hans-Christian Egtvedt <hans-christian.egtvedt@...el.com>,
Ralf Baechle <ralf@...ux-mips.org>,
Russell King <linux@....linux.org.uk>,
Paul Mackerras <paulus@...ba.org>,
Heiko Carstens <heiko.carstens@...ibm.com>,
Paul Mundt <lethal@...ux-sh.org>
Subject: Re: [PATCH 0/5 v3] rcu: Fix some rcu uses in extended quiescent
state
On Sat, Sep 17, 2011 at 04:39:57PM +0200, Frederic Weisbecker wrote:
> This is an updated set that tries to fix some illegal uses
> of RCU that are made in extended quiescent state, ie: between
> calls of rcu_enter_nohz() and rcu_exit_nohz().
Thank you, Frederic!
I have queued these, and have started testing them.
Thanx, Paul
> Changes:
>
> - Restructuring of the changes in nohz core code. We need to handle
> the case when the tick is not stopped from tick_nohz_idle_enter() but
> from the irq exit path when the irq happens in idle loop. This should
> fix the RCU stall Paul detected after applying my patches.
> The first patch sort of overlaps the simple scope of fixes for RCU
> but I believe it makes the code slightly more simple and optimized
> in the end.
>
> - Rename tick_nohz_enter_idle() into tick_nohz_idle_enter() (same for
> exit). That to stay coherent with usual naming (tips from Peterz on
> the nohz cpuset patchset that needs similar changes).
>
> - Remove the rcu_ext_qs parameter from tick_nohz_idle_exit(). We only
> need it in tick_nohz_idle_enter() because as long as we call
> rcu_enter_nohz() explicitly, we really want to pair it with an explicit
> call to rcu_exit_nohz() (and vice versa), to avoid confusion.
>
> There are still the powerpc fixes that are not included. But I would
> like to first get this set merged in Paul's tree before focusing on
> the powerpc reported issues.
>
> Frederic Weisbecker (5):
> nohz: Seperate out irq exit and idle loop dyntick logic
> nohz: Allow rcu extended state handling seperately from tick stop
> x86: Enter rcu extended qs after idle notifier call
> x86: Call idle notifier after irq_enter()
> rcu: Fix early call to rcu_irq_exit()
>
> arch/arm/kernel/process.c | 4 +-
> arch/avr32/kernel/process.c | 4 +-
> arch/blackfin/kernel/process.c | 4 +-
> arch/microblaze/kernel/process.c | 4 +-
> arch/mips/kernel/process.c | 4 +-
> arch/powerpc/kernel/idle.c | 4 +-
> arch/powerpc/platforms/iseries/setup.c | 8 +-
> arch/s390/kernel/process.c | 4 +-
> arch/sh/kernel/idle.c | 4 +-
> arch/sparc/kernel/process_64.c | 4 +-
> arch/tile/kernel/process.c | 4 +-
> arch/um/kernel/process.c | 4 +-
> arch/unicore32/kernel/process.c | 4 +-
> arch/x86/kernel/apic/apic.c | 6 +-
> arch/x86/kernel/apic/io_apic.c | 2 +-
> arch/x86/kernel/cpu/mcheck/mce.c | 2 +-
> arch/x86/kernel/cpu/mcheck/therm_throt.c | 2 +-
> arch/x86/kernel/cpu/mcheck/threshold.c | 2 +-
> arch/x86/kernel/irq.c | 6 +-
> arch/x86/kernel/process_32.c | 4 +-
> arch/x86/kernel/process_64.c | 9 ++-
> include/linux/tick.h | 12 ++-
> kernel/softirq.c | 4 +-
> kernel/time/tick-sched.c | 110 ++++++++++++++++++++----------
> 24 files changed, 131 insertions(+), 84 deletions(-)
>
> --
> 1.7.5.4
>
--
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/
Powered by blists - more mailing lists