From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ip4-83-240-67-251.cust.nbox.cz ([83.240.67.251]:51320 "EHLO ip4-83-240-18-248.cust.nbox.cz" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1752915AbbEDMiI (ORCPT ); Mon, 4 May 2015 08:38:08 -0400 From: Jiri Slaby To: stable@vger.kernel.org Cc: Calvin Owens , "Paul E. McKenney" , Mike Galbraith , Jiri Slaby Subject: [patch added to the 3.12 stable tree] ksoftirqd: Enable IRQs and call cond_resched() before poking RCU Date: Mon, 4 May 2015 14:37:54 +0200 Message-Id: <1430743082-6957-63-git-send-email-jslaby@suse.cz> In-Reply-To: <1430743082-6957-1-git-send-email-jslaby@suse.cz> References: <1430743082-6957-1-git-send-email-jslaby@suse.cz> Sender: stable-owner@vger.kernel.org List-ID: From: Calvin Owens This patch has been added to the 3.12 stable tree. If you have any objections, please let us know. =============== commit 28423ad283d5348793b0c45cc9b1af058e776fd6 upstream. While debugging an issue with excessive softirq usage, I encountered the following note in commit 3e339b5dae24a706 ("softirq: Use hotplug thread infrastructure"): [ paulmck: Call rcu_note_context_switch() with interrupts enabled. ] ...but despite this note, the patch still calls RCU with IRQs disabled. This seemingly innocuous change caused a significant regression in softirq CPU usage on the sending side of a large TCP transfer (~1 GB/s): when introducing 0.01% packet loss, the softirq usage would jump to around 25%, spiking as high as 50%. Before the change, the usage would never exceed 5%. Moving the call to rcu_note_context_switch() after the cond_sched() call, as it was originally before the hotplug patch, completely eliminated this problem. Signed-off-by: Calvin Owens Signed-off-by: Paul E. McKenney Signed-off-by: Mike Galbraith Signed-off-by: Jiri Slaby --- kernel/softirq.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/kernel/softirq.c b/kernel/softirq.c index d7d498d8cc4f..b331c8756543 100644 --- a/kernel/softirq.c +++ b/kernel/softirq.c @@ -772,9 +772,13 @@ static void run_ksoftirqd(unsigned int cpu) local_irq_disable(); if (local_softirq_pending()) { __do_softirq(); - rcu_note_context_switch(cpu); local_irq_enable(); cond_resched(); + + preempt_disable(); + rcu_note_context_switch(cpu); + preempt_enable(); + return; } local_irq_enable(); -- 2.3.5