2013-04-26 07:57:39

by Li Fei

[permalink] [raw]
Subject: [PATCH] x86: eliminate irq_mis_count counted in arch_irq_stat


With current implementation, kstat_cpu(cpu).irqs_sum is also
increased in case of irq_mis_count increment.
So there is no need to count irq_mis_count in arch_irq_stat,
otherwise irq_mis_count will be counted twice in the sum of
/proc/stat.

Signed-off-by: Liu Chuansheng <[email protected]>
Signed-off-by: Li Fei <[email protected]>
---
arch/x86/kernel/irq.c | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index e4595f1..84b7789 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -165,10 +165,6 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
u64 arch_irq_stat(void)
{
u64 sum = atomic_read(&irq_err_count);
-
-#ifdef CONFIG_X86_IO_APIC
- sum += atomic_read(&irq_mis_count);
-#endif
return sum;
}

--
1.7.4.1



2013-04-26 08:19:23

by Ingo Molnar

[permalink] [raw]
Subject: Re: [PATCH] x86: eliminate irq_mis_count counted in arch_irq_stat


* Li Fei <[email protected]> wrote:

>
> With current implementation, kstat_cpu(cpu).irqs_sum is also
> increased in case of irq_mis_count increment.
> So there is no need to count irq_mis_count in arch_irq_stat,
> otherwise irq_mis_count will be counted twice in the sum of
> /proc/stat.
>
> Signed-off-by: Liu Chuansheng <[email protected]>
> Signed-off-by: Li Fei <[email protected]>

The patch looks good, but this is not a valid signoff sequence. If Liu
Chuansheng wrote this patch then his From: line should be included. If he
acked it, then it should be added as Acked-by.

Thanks,

Ingo

2013-04-26 08:44:59

by Li Fei

[permalink] [raw]
Subject: RE: [PATCH] x86: eliminate irq_mis_count counted in arch_irq_stat

> > Signed-off-by: Liu Chuansheng <[email protected]>
> > Signed-off-by: Li Fei <[email protected]>
>
> The patch looks good, but this is not a valid signoff sequence. If Liu
> Chuansheng wrote this patch then his From: line should be included. If he
> acked it, then it should be added as Acked-by.
>
Hello Ingo,
Thanks for your quick feedback.

I add Chuansheng's signature basing on:
1> Chuansheng found such hole in code;
2> After discussion with Chuansheng, I wrote the patch;
3> Chuansheng also acked it.

Do you think it's OK to add Chuansheng's signature in such case?
Or better suggestion?

Thanks in advance!
Fei

> Thanks,
>
> Ingo

2013-04-26 09:33:34

by Ingo Molnar

[permalink] [raw]
Subject: Re: [PATCH] x86: eliminate irq_mis_count counted in arch_irq_stat


* Li, Fei <[email protected]> wrote:

> > > Signed-off-by: Liu Chuansheng <[email protected]>
> > > Signed-off-by: Li Fei <[email protected]>
> >
> > The patch looks good, but this is not a valid signoff sequence. If Liu
> > Chuansheng wrote this patch then his From: line should be included. If he
> > acked it, then it should be added as Acked-by.
> >
> Hello Ingo,
> Thanks for your quick feedback.
>
> I add Chuansheng's signature basing on:
> 1> Chuansheng found such hole in code;
> 2> After discussion with Chuansheng, I wrote the patch;
> 3> Chuansheng also acked it.
>
> Do you think it's OK to add Chuansheng's signature in such case?
> Or better suggestion?

Signed-off-by has a specific meaning, it typically means someone authored
or as a maintainer forwarded the code (which did not happen here) - see
the details in Documentation/SubmittingPatches, under this section:

12) Sign your work

If Chuansheng found the bug, there's a Reported-by tag. If Chuansheng
approves of the fix patch then there's Acked-by for that. If he tested it,
then for that there's the Tested-by tag. You can also add several of those
tags.

Thanks,

Ingo

2013-04-26 12:51:48

by Li Fei

[permalink] [raw]
Subject: [PATCH V2] x86: eliminate irq_mis_count counted in arch_irq_stat


With current implementation, kstat_cpu(cpu).irqs_sum is also
increased in case of irq_mis_count increment.
So there is no need to count irq_mis_count in arch_irq_stat,
otherwise irq_mis_count will be counted twice in the sum of
/proc/stat.

Reported-by: Liu Chuansheng <[email protected]>
Signed-off-by: Li Fei <[email protected]>
Acked-by: Liu Chuansheng <[email protected]>
---
arch/x86/kernel/irq.c | 4 ----
1 files changed, 0 insertions(+), 4 deletions(-)

diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index e4595f1..84b7789 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -165,10 +165,6 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
u64 arch_irq_stat(void)
{
u64 sum = atomic_read(&irq_err_count);
-
-#ifdef CONFIG_X86_IO_APIC
- sum += atomic_read(&irq_mis_count);
-#endif
return sum;
}

--
1.7.4.1


Subject: [tip:perf/urgent] x86: Eliminate irq_mis_count counted in arch_irq_stat

Commit-ID: f7b0e1055574ce06ab53391263b4e205bf38daf3
Gitweb: http://git.kernel.org/tip/f7b0e1055574ce06ab53391263b4e205bf38daf3
Author: Li Fei <[email protected]>
AuthorDate: Fri, 26 Apr 2013 20:50:11 +0800
Committer: Ingo Molnar <[email protected]>
CommitDate: Tue, 30 Apr 2013 10:56:37 +0200

x86: Eliminate irq_mis_count counted in arch_irq_stat

With the current implementation, kstat_cpu(cpu).irqs_sum is also
increased in case of irq_mis_count increment.

So there is no need to count irq_mis_count in arch_irq_stat,
otherwise irq_mis_count will be counted twice in the sum of
/proc/stat.

Reported-by: Liu Chuansheng <[email protected]>
Signed-off-by: Li Fei <[email protected]>
Acked-by: Liu Chuansheng <[email protected]>
Cc: [email protected]
Cc: [email protected]
Link: http://lkml.kernel.org/r/[email protected]
Cc: <[email protected]>
Signed-off-by: Ingo Molnar <[email protected]>
---
arch/x86/kernel/irq.c | 4 ----
1 file changed, 4 deletions(-)

diff --git a/arch/x86/kernel/irq.c b/arch/x86/kernel/irq.c
index e4595f1..84b7789 100644
--- a/arch/x86/kernel/irq.c
+++ b/arch/x86/kernel/irq.c
@@ -165,10 +165,6 @@ u64 arch_irq_stat_cpu(unsigned int cpu)
u64 arch_irq_stat(void)
{
u64 sum = atomic_read(&irq_err_count);
-
-#ifdef CONFIG_X86_IO_APIC
- sum += atomic_read(&irq_mis_count);
-#endif
return sum;
}