From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1754525Ab0AUE0r (ORCPT ); Wed, 20 Jan 2010 23:26:47 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1754161Ab0AUE0o (ORCPT ); Wed, 20 Jan 2010 23:26:44 -0500 Received: from kroah.org ([198.145.64.141]:60174 "EHLO coco.kroah.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754208Ab0AUEXD (ORCPT ); Wed, 20 Jan 2010 23:23:03 -0500 X-Mailbox-Line: From gregkh@mini.kroah.org Wed Jan 20 20:18:35 2010 Message-Id: <20100121041835.475421852@mini.kroah.org> User-Agent: quilt/0.48-1 Date: Wed, 20 Jan 2010 20:15:44 -0800 From: Greg KH To: linux-kernel@vger.kernel.org, stable@kernel.org Cc: stable-review@kernel.org, torvalds@linux-foundation.org, akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk, "David S. Miller" , Peter Zijlstra , Mike Galbraith , Ingo Molnar Subject: [23/30] sched: Fix cpu_clock() in NMIs, on !CONFIG_HAVE_UNSTABLE_SCHED_CLOCK In-Reply-To: <20100121041852.GA9656@kroah.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org 2.6.32-stable review patch. If anyone has any objections, please let us know. ------------------ From: David Miller commit b9f8fcd55bbdb037e5332dbdb7b494f0b70861ac upstream. Relax stable-sched-clock architectures to not save/disable/restore hardirqs in cpu_clock(). The background is that I was trying to resolve a sparc64 perf issue when I discovered this problem. On sparc64 I implement pseudo NMIs by simply running the kernel at IRQ level 14 when local_irq_disable() is called, this allows performance counter events to still come in at IRQ level 15. This doesn't work if any code in an NMI handler does local_irq_save() or local_irq_disable() since the "disable" will kick us back to cpu IRQ level 14 thus letting NMIs back in and we recurse. The only path which that does that in the perf event IRQ handling path is the code supporting frequency based events. It uses cpu_clock(). cpu_clock() simply invokes sched_clock() with IRQs disabled. And that's a fundamental bug all on it's own, particularly for the HAVE_UNSTABLE_SCHED_CLOCK case. NMIs can thus get into the sched_clock() code interrupting the local IRQ disable code sections of it. Furthermore, for the not-HAVE_UNSTABLE_SCHED_CLOCK case, the IRQ disabling done by cpu_clock() is just pure overhead and completely unnecessary. So the core problem is that sched_clock() is not NMI safe, but we are invoking it from NMI contexts in the perf events code (via cpu_clock()). A less important issue is the overhead of IRQ disabling when it isn't necessary in cpu_clock(). CONFIG_HAVE_UNSTABLE_SCHED_CLOCK architectures are not affected by this patch. Signed-off-by: David S. Miller Acked-by: Peter Zijlstra Cc: Mike Galbraith LKML-Reference: <20091213.182502.215092085.davem@davemloft.net> Signed-off-by: Ingo Molnar Signed-off-by: Greg Kroah-Hartman --- kernel/sched_clock.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-) --- a/kernel/sched_clock.c +++ b/kernel/sched_clock.c @@ -236,6 +236,18 @@ void sched_clock_idle_wakeup_event(u64 d } EXPORT_SYMBOL_GPL(sched_clock_idle_wakeup_event); +unsigned long long cpu_clock(int cpu) +{ + unsigned long long clock; + unsigned long flags; + + local_irq_save(flags); + clock = sched_clock_cpu(cpu); + local_irq_restore(flags); + + return clock; +} + #else /* CONFIG_HAVE_UNSTABLE_SCHED_CLOCK */ void sched_clock_init(void) @@ -251,17 +263,12 @@ u64 sched_clock_cpu(int cpu) return sched_clock(); } -#endif /* CONFIG_HAVE_UNSTABLE_SCHED_CLOCK */ unsigned long long cpu_clock(int cpu) { - unsigned long long clock; - unsigned long flags; + return sched_clock_cpu(cpu); +} - local_irq_save(flags); - clock = sched_clock_cpu(cpu); - local_irq_restore(flags); +#endif /* CONFIG_HAVE_UNSTABLE_SCHED_CLOCK */ - return clock; -} EXPORT_SYMBOL_GPL(cpu_clock);