[PATCH] x86: perf: fix incorrect use of do_div() in nmi warning

[PATCH] x86: perf: fix incorrect use of do_div() in nmi warning

[Dave Hansen]

From: Dave Hansen <dave.hansen@linux.intel.com>

This fixes a bug present in 3.10 and introduced here:

commit 2ab00456ea8a0d79acb1390659b98416111880b2
Author: Dave Hansen <dave.hansen@linux.intel.com>
Date:   Fri Jun 21 08:51:35 2013 -0700

    x86: Warn when NMI handlers take large amounts of time

I completely botched understanding the calling conventions of
do_div().  I assumed that do_div() returned the result instead
of realizing that it modifies its argument and returns a
remainder.  The side-effect from this would be bogus numbers
for the "msecs" value in the warning messages:

	INFO: NMI handler (perf_event_nmi_handler) took too long to run: 0.114 msecs

Note, there was a second fix posted by Stephane Eranian for
a separate patch which I also botched:

	http://lkml.kernel.org/r/20130704223010.GA30625@quad

Both of these fixes need to get pulled in to Linus's tree and
the 3.10 stable tree.

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: stable@vger.kernel.org
Cc: x86@kernel.org
---

 linux.git-davehans/arch/x86/kernel/nmi.c |    7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff -puN arch/x86/kernel/nmi.c~dave-is-stupid-do_div arch/x86/kernel/nmi.c
--- linux.git/arch/x86/kernel/nmi.c~dave-is-stupid-do_div	2013-07-08 13:10:55.498656504 -0700
+++ linux.git-davehans/arch/x86/kernel/nmi.c	2013-07-08 13:10:55.501656637 -0700
@@ -111,7 +111,7 @@ static int __kprobes nmi_handle(unsigned
 	 */
 	list_for_each_entry_rcu(a, &desc->head, list) {
 		u64 before, delta, whole_msecs;
-		int decimal_msecs, thishandled;
+		int remainder_ns, decimal_msecs, thishandled;
 
 		before = local_clock();
 		thishandled = a->handler(type, regs);
@@ -123,8 +123,9 @@ static int __kprobes nmi_handle(unsigned
 			continue;
 
 		nmi_longest_ns = delta;
-		whole_msecs = do_div(delta, (1000 * 1000));
-		decimal_msecs = do_div(delta, 1000) % 1000;
+		whole_msecs = delta;
+		remainder_ns = do_div(whole_msecs, (1000 * 1000));
+		decimal_msecs = remainder_ns / 1000;
 		printk_ratelimited(KERN_INFO
 			"INFO: NMI handler (%ps) took too long to run: "
 			"%lld.%03d msecs\n", a->handler, whole_msecs,
_

Re: [PATCH] x86: perf: fix incorrect use of do_div() in nmi warning

[Ingo Molnar]

* Dave Hansen <dave@sr71.net> wrote:

> 
> From: Dave Hansen <dave.hansen@linux.intel.com>
> 
> This fixes a bug present in 3.10 and introduced here:
> 
> commit 2ab00456ea8a0d79acb1390659b98416111880b2
> Author: Dave Hansen <dave.hansen@linux.intel.com>
> Date:   Fri Jun 21 08:51:35 2013 -0700
> 
>     x86: Warn when NMI handlers take large amounts of time
> 
> I completely botched understanding the calling conventions of
> do_div().  I assumed that do_div() returned the result instead
> of realizing that it modifies its argument and returns a
> remainder.  The side-effect from this would be bogus numbers
> for the "msecs" value in the warning messages:
> 
> 	INFO: NMI handler (perf_event_nmi_handler) took too long to run: 0.114 msecs
> 
> Note, there was a second fix posted by Stephane Eranian for
> a separate patch which I also botched:
> 
> 	http://lkml.kernel.org/r/20130704223010.GA30625@quad
> 
> Both of these fixes need to get pulled in to Linus's tree and
> the 3.10 stable tree.

This is a brand new v3.11 commit.

It should not be included in any -stable tree, is it?

Thanks,

	Ingo

[tip:perf/urgent] perf/x86: Fix incorrect use of do_div() in NMI warning

[tip-bot for Dave Hansen]

Commit-ID:  baf64b85445546a38b44052d71782dfe7531e350
Gitweb:     http://git.kernel.org/tip/baf64b85445546a38b44052d71782dfe7531e350
Author:     Dave Hansen <dave.hansen@linux.intel.com>
AuthorDate: Mon, 8 Jul 2013 14:44:04 -0700
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Fri, 12 Jul 2013 14:13:04 +0200

perf/x86: Fix incorrect use of do_div() in NMI warning

I completely botched understanding the calling conventions of
do_div().  I assumed that do_div() returned the result instead
of realizing that it modifies its argument and returns a
remainder.  The side-effect from this would be bogus numbers
for the "msecs" value in the warning messages:

	INFO: NMI handler (perf_event_nmi_handler) took too long to run: 0.114 msecs

Note, there was a second fix posted by Stephane Eranian for
a separate patch which I also botched:

	http://lkml.kernel.org/r/20130704223010.GA30625@quad

Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Dave Hansen <dave@sr71.net>
Link: http://lkml.kernel.org/r/20130708214404.B0B6EA66@viggo.jf.intel.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 arch/x86/kernel/nmi.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kernel/nmi.c b/arch/x86/kernel/nmi.c
index 0920212..ba77ebc 100644
--- a/arch/x86/kernel/nmi.c
+++ b/arch/x86/kernel/nmi.c
@@ -111,7 +111,7 @@ static int __kprobes nmi_handle(unsigned int type, struct pt_regs *regs, bool b2
 	 */
 	list_for_each_entry_rcu(a, &desc->head, list) {
 		u64 before, delta, whole_msecs;
-		int decimal_msecs, thishandled;
+		int remainder_ns, decimal_msecs, thishandled;
 
 		before = local_clock();
 		thishandled = a->handler(type, regs);
@@ -123,8 +123,9 @@ static int __kprobes nmi_handle(unsigned int type, struct pt_regs *regs, bool b2
 			continue;
 
 		nmi_longest_ns = delta;
-		whole_msecs = do_div(delta, (1000 * 1000));
-		decimal_msecs = do_div(delta, 1000) % 1000;
+		whole_msecs = delta;
+		remainder_ns = do_div(whole_msecs, (1000 * 1000));
+		decimal_msecs = remainder_ns / 1000;
 		printk_ratelimited(KERN_INFO
 			"INFO: NMI handler (%ps) took too long to run: "
 			"%lld.%03d msecs\n", a->handler, whole_msecs,

Re: [PATCH] x86: perf: fix incorrect use of do_div() in nmi warning

[Dave Hansen]

On 07/12/2013 05:08 AM, Ingo Molnar wrote:
>> > Note, there was a second fix posted by Stephane Eranian for
>> > a separate patch which I also botched:
>> > 
>> > 	http://lkml.kernel.org/r/20130704223010.GA30625@quad
>> > 
>> > Both of these fixes need to get pulled in to Linus's tree and
>> > the 3.10 stable tree.
> This is a brand new v3.11 commit.
> 
> It should not be included in any -stable tree, is it?

Ingo, you are correct.  The commit only appeared in Linus's tree
post-3.10.  It does not need to be in any stable trees.