Attachment 156165 Details for Bug 242422 – rostedt.patch

rostedt.patch

rostedt.patch (text/plain), 8.04 KB, created by IBM Bug Proxy on 2007-06-05 01:15:45 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: IBM Bug Proxy

Created: 2007-06-05 01:15:45 UTC

Size: 8.04 KB

patch

obsolete

>From rostedt@goodmis.org Mon Jun  4 17:56:45 2007
>Return-Path: <tglx@linutronix.de>
>Received: from imap.linux.ibm.com ([unix socket]) by imap.linux.ibm.com
>	(Cyrus v2.3.7-Invoca-RPM-2.3.7-7) with LMTPA; Mon, 04 Jun 2007 17:56:45
>	-0400
>X-Sieve: CMU Sieve 2.3
>Received: by imap.linux.ibm.com (Postfix, from userid 101) id EB9091910153;
>	Mon,  4 Jun 2007 17:56:45 -0400 (EDT)
>X-Spam-TestScore: none
>X-Spam-TokenSummary: Bayes not run.
>X-Spam-Checker-Version: SpamAssassin 3.1.7 (2006-10-05) on
>	imap.linux.ibm.com
>X-Spam-Level: 
>X-Spam-Status: No, score=0.0 required=5.0 tests=none autolearn=disabled 
>	version=3.1.7
>X-Spam-Relay-Country: US ** US US US ** US US DE DE US US
>Received: from smtp.linux.ibm.com (smtp.linux.ibm.com [9.26.4.197]) by
>	imap.linux.ibm.com (Postfix) with ESMTP id 891E9191013E for
>	<johnstul@imap.linux.ibm.com>; Mon,  4 Jun 2007 17:56:39 -0400 (EDT)
>Received: from localhost (localhost.localdomain [127.0.0.1]) by
>	smtp.linux.ibm.com (Postfix) with ESMTP id 27EA8C03E for
>	<johnstul@linux.ibm.com>; Mon,  4 Jun 2007 17:56:39 -0400 (EDT)
>X-Virus-Scanned: amavisd-new at linux.ibm.com
>Received: from VMSDVMA.POK.IBM.COM (vmsdvma.pok.ibm.com [9.56.231.65]) by
>	smtp.linux.ibm.com (Postfix) with ESMTP id E96A8C040 for
>	<johnstul@linux.ibm.com>; Mon,  4 Jun 2007 17:56:38 -0400 (EDT)
>Received:  by VMSDVMA.POK.IBM.COM (IBM VM SMTP Level 520) via spool with
>	SMTP id 2456 ; Mon, 04 Jun 2007 17:55:03 EDT
>Received: by vmsdvma.vnet.ibm.com (xagent2 5.8.6) via xagsmtp3 with spool
>	id 5678 for johnstul@linux.vnet.ibm.com; Mon, 04 Jun 2007 17:55:03 -0400
>	(EDT)
>Received: from d01relay03.pok.ibm.com [9.56.227.235] by VMSDVMA.POK.IBM.COM
>	(IBM VM SMTP Level 520) via TCP with ESMTP ; Mon, 04 Jun 2007 17:55:00 EDT
>Received: from d01av04.pok.ibm.com (d01av04.pok.ibm.com [9.56.224.64]) by
>	d01relay03.pok.ibm.com (8.13.8/8.13.8/NCO v8.3) with ESMTP id
>	l54LuQAT257558 for <johnstul@us.ibm.com>; Mon, 4 Jun 2007 17:56:28 -0400
>Received: from d01av04.pok.ibm.com (loopback [127.0.0.1]) by
>	d01av04.pok.ibm.com (8.12.11.20060308/8.13.3) with ESMTP id l54LuQbF030438
>	for <johnstul@us.ibm.com>; Mon, 4 Jun 2007 17:56:26 -0400
>Received: from d01as03.pok.ibm.com (d01as03 [9.56.224.40]) by
>	d01av04.pok.ibm.com (8.12.11.20060308/8.12.11) with ESMTP id l54LuQpO030435
>	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for
>	<johnstul@us.ibm.com>; Mon, 4 Jun 2007 17:56:26 -0400
>Received: from e6.ny.us.ibm.com (e6.pok.ibm.com [9.56.232.146]) by
>	d01as03.pok.ibm.com (8.12.11.20060308/8.12.11) with ESMTP id l54LuOeq014708
>	(version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for
>	<johnstul@us.ibm.com>; Mon, 4 Jun 2007 17:56:25 -0400
>Received: from mail.tglx.de (www.osadl.org [213.239.205.134]) by
>	e6.ny.us.ibm.com (8.13.8/8.13.8) with ESMTP id l54LvSPH023798 for
>	<johnstul@us.ibm.com>; Mon, 4 Jun 2007 17:57:28 -0400
>Received: from [127.0.0.1] (debian [213.239.205.147]) by mail.tglx.de
>	(Postfix) with ESMTP id D847665C065 for <johnstul@us.ibm.com>; Mon, 4 Jun
>	2007 23:56:13 +0200 (CEST)
>Received: from ms-smtp-05.nyroc.rr.com (ms-smtp-05.nyroc.rr.com
>	[24.24.2.59]) by mail.tglx.de (Postfix) with ESMTP id 89CE665C065 for
>	<tglx@linutronix.de>; Mon, 4 Jun 2007 23:01:02 +0200 (CEST)
>Received: from [192.168.23.10] (cpe-24-94-51-176.stny.res.rr.com
>	[24.94.51.176]) by ms-smtp-05.nyroc.rr.com (8.13.6/8.13.6) with ESMTP id
>	l54L0m7g003748; Mon, 4 Jun 2007 17:00:48 -0400 (EDT)
>From: Steven Rostedt <rostedt@goodmis.org>
>To: john stultz <johnstul@us.ibm.com>
>Cc: Thomas Gleixner <tglx@linutronix.de>, Ingo Molnar <mingo@elte.hu>,  Oleg Nesterov <oleg@tv-sign.ru>, sshtylyov@ru.mvista.com
>Content-Type: text/plain
>Message-Id: <1180990847.18444.9.camel@localhost.localdomain>
>Mime-Version: 1.0
>X-Virus-Scanned: Symantec AntiVirus Scan Engine
>Subject: [HACK] see if this fixes the softirq warn-on
>Resent-From: Thomas Gleixner <tglx@linutronix.de>
>Resent-To: john stultz <johnstul@us.ibm.com>
>Date: Mon, 04 Jun 2007 23:56:13 +0200
>X-Mailer: Evolution 2.10.1 (2.10.1-4.fc7) 
>Resent-Message-Id: <20070604215613.D847665C065@mail.tglx.de>
>Resent-Date: Mon,  4 Jun 2007 23:56:13 +0200 (CEST)
>X-Xagent-From: rostedt@goodmis.org
>X-Xagent-To: johnstul@linux.vnet.ibm.com
>X-Xagent-Gateway: vmsdvma.vnet.ibm.com (XAGENTU at VMSDVMA)
>X-Evolution-Source: imap://johnstul@imap.linux.ibm.com/
>Content-Transfer-Encoding: 8bit
>
>John,
>
>here's my hack to see if this solves that tasklet problem.
>
>For those others that don't know the issue:
>
>We've been seeing a lot of WARN_ON coming from the tasklet code.
>Specifically:
>
>                if (!test_and_clear_bit(TASKLET_STATE_SCHED, &t->state))
>                        WARN_ON(1);
>
>
>Looking into this, one place that might be a problem is the
>tasklet_kill.  It seems that with the sleeps there, and the fact that
>drivers seems to use this quite a bit, there's really no guarantee that
>after a the tasklet has been waited on that it's not scheduled again
>before it clears the TASKLET_STATE_SCHED bit.
>
>So here's a hack that tries to fix this.
>
>Comments?
>
>John, does this solve the issue on your test boxes?
>
>Signed-off-by: Steven Rostedt <srostedt@redhat.com>
>
>Index: linux-2.6.21-rt9/kernel/softirq.c
>===================================================================
>--- linux-2.6.21-rt9.orig/kernel/softirq.c
>+++ linux-2.6.21-rt9/kernel/softirq.c
>@@ -535,10 +535,14 @@ struct tasklet_head
> 	struct tasklet_struct *list;
> };
>
>-/* Some compilers disobey section attribute on statics when not
>-   initialized -- RR */
>-static DEFINE_PER_CPU(struct tasklet_head, tasklet_vec) = { NULL };
>-static DEFINE_PER_CPU(struct tasklet_head, tasklet_hi_vec) = { NULL };
>+/*
>+ * Initialize to something other than NULL so we can test
>+ * if a tasklet is on the list or not;
>+ */
>+#define TASKLET_LIST_END ((void*)0x101)
>+
>+static DEFINE_PER_CPU(struct tasklet_head, tasklet_vec) = { TASKLET_LIST_END };
>+static DEFINE_PER_CPU(struct tasklet_head, tasklet_hi_vec) = { TASKLET_LIST_END };
>
> static void inline
> __tasklet_common_schedule(struct tasklet_struct *t, struct tasklet_head *head, unsigned int nr)
>@@ -599,9 +603,14 @@ __tasklet_action(struct softirq_action *
> {
> 	int loops = 1000000;
>
>-	while (list) {
>+	while (list != TASKLET_LIST_END) {
> 		struct tasklet_struct *t = list;
>
>+		if (!list) {
>+			WARN_ON(1);
>+			break;
>+		}
>+
> 		list = list->next;
> 		/*
> 		 * Should always succeed - after a tasklist got on the
>@@ -674,7 +683,7 @@ static void tasklet_action(struct softir
>
> 	local_irq_disable();
> 	list = __get_cpu_var(tasklet_vec).list;
>-	__get_cpu_var(tasklet_vec).list = NULL;
>+	__get_cpu_var(tasklet_vec).list = TASKLET_LIST_END;
> 	local_irq_enable();
>
> 	__tasklet_action(a, list);
>@@ -686,7 +695,7 @@ static void tasklet_hi_action(struct sof
>
> 	local_irq_disable();
> 	list = __get_cpu_var(tasklet_hi_vec).list;
>-	__get_cpu_var(tasklet_hi_vec).list = NULL;
>+	__get_cpu_var(tasklet_hi_vec).list = TASKLET_LIST_END;
> 	local_irq_enable();
>
> 	__tasklet_action(a, list);
>@@ -709,13 +718,19 @@ void tasklet_kill(struct tasklet_struct 
> 	if (in_interrupt())
> 		printk("Attempt to kill tasklet from interrupt\n");
>
>+again:
> 	while (test_and_set_bit(TASKLET_STATE_SCHED, &t->state)) {
> 		do
> 			msleep(1);
> 		while (test_bit(TASKLET_STATE_SCHED, &t->state));
> 	}
> 	tasklet_unlock_wait(t);
>-	clear_bit(TASKLET_STATE_SCHED, &t->state);
>+	if (test_bit(TASKLET_STATE_SCHED, &t->state)) {
>+		/* only clear the bit if the tasklet is not added back to the list */
>+		if (t->next || test_bit(TASKLET_STATE_RUN, &t->state))
>+			goto again; /* still scheduled?? */
>+		clear_bit(TASKLET_STATE_SCHED, &t->state);
>+	}
> }
>
> EXPORT_SYMBOL(tasklet_kill);
>@@ -873,12 +888,12 @@ void takeover_tasklets(unsigned int cpu)
> 	/* Find end, append list for that CPU. */
> 	for (i = &__get_cpu_var(tasklet_vec).list; *i; i = &(*i)->next);
> 	*i = per_cpu(tasklet_vec, cpu).list;
>-	per_cpu(tasklet_vec, cpu).list = NULL;
>+	per_cpu(tasklet_vec, cpu).list = TASKLET_LIST_END;
> 	raise_softirq_irqoff(TASKLET_SOFTIRQ);
>
> 	for (i = &__get_cpu_var(tasklet_hi_vec).list; *i; i = &(*i)->next);
> 	*i = per_cpu(tasklet_hi_vec, cpu).list;
>-	per_cpu(tasklet_hi_vec, cpu).list = NULL;
>+	per_cpu(tasklet_hi_vec, cpu).list = TASKLET_LIST_END;
> 	raise_softirq_irqoff(HI_SOFTIRQ);
>
> 	local_irq_enable();
>
>
>

Actions: View

Attachments on bug 242422: 156165 | 156301