2009-06-10 18:48:53

by Jaswinder Singh Rajput

Subject: [PATCH -tip] x86: perf_counter AMD supports performance monitoring for K7

Signed-off-by: Jaswinder Singh Rajput <[email protected]>
---
arch/x86/kernel/cpu/perf_counter.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
index 240ca56..5d56742 100644
--- a/arch/x86/kernel/cpu/perf_counter.c
+++ b/arch/x86/kernel/cpu/perf_counter.c
@@ -1436,6 +1436,10 @@ static int intel_pmu_init(void)

static int amd_pmu_init(void)
{
+ /* performance-monitoring supported from K7 */
+ if (boot_cpu_data.x86 < 7)
+ return -ENODEV;
+
x86_pmu = amd_pmu;

switch (boot_cpu_data.x86) {
--
1.6.1.1

2009-06-11 00:13:53

by Ingo Molnar

Subject: Re: [PATCH -tip] x86: perf_counter AMD supports performance monitoring for K7

* Jaswinder Singh Rajput <[email protected]> wrote:

>
> Signed-off-by: Jaswinder Singh Rajput <[email protected]>
> ---
> arch/x86/kernel/cpu/perf_counter.c | 4 ++++
> 1 files changed, 4 insertions(+), 0 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
> index 240ca56..5d56742 100644
> --- a/arch/x86/kernel/cpu/perf_counter.c
> +++ b/arch/x86/kernel/cpu/perf_counter.c
> @@ -1436,6 +1436,10 @@ static int intel_pmu_init(void)
>
> static int amd_pmu_init(void)
> {
> + /* performance-monitoring supported from K7 */
> + if (boot_cpu_data.x86 < 7)
> + return -ENODEV;
> +
> x86_pmu = amd_pmu;
>
> switch (boot_cpu_data.x86) {

A commit under Linux consists of three parts:

1) the title
2) the changlog
3) the body

This sloppy patch of yours manages to get _all_ three wrong:

- the title is wrong (it is incorrect)
- the changelog is wrong (it is missing)
- the body is wrong (it is broken - K7 != family 7)

What the heck is going on here?

Ingo

2009-06-11 05:33:30

by Jaswinder Singh Rajput

Subject: Re: [PATCH -tip] x86: perf_counter AMD supports performance monitoring for K7

On Thu, 2009-06-11 at 02:13 +0200, Ingo Molnar wrote:
> * Jaswinder Singh Rajput <[email protected]> wrote:
>
> >
> > Signed-off-by: Jaswinder Singh Rajput <[email protected]>
> > ---
> > arch/x86/kernel/cpu/perf_counter.c | 4 ++++
> > 1 files changed, 4 insertions(+), 0 deletions(-)
> >
> > diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
> > index 240ca56..5d56742 100644
> > --- a/arch/x86/kernel/cpu/perf_counter.c
> > +++ b/arch/x86/kernel/cpu/perf_counter.c
> > @@ -1436,6 +1436,10 @@ static int intel_pmu_init(void)
> >
> > static int amd_pmu_init(void)
> > {
> > + /* performance-monitoring supported from K7 */
> > + if (boot_cpu_data.x86 < 7)
> > + return -ENODEV;
> > +
> > x86_pmu = amd_pmu;
> >
> > switch (boot_cpu_data.x86) {
>
> A commit under Linux consists of three parts:
>
> 1) the title
> 2) the changlog
> 3) the body
>
> This sloppy patch of yours manages to get _all_ three wrong:
>
> - the title is wrong (it is incorrect)
> - the changelog is wrong (it is missing)
> - the body is wrong (it is broken - K7 != family 7)
>

oops, sorry here is updated patch :

[PATCH -tip] x86: perf_counter check AMD performance monitoring support

AMD supports performance monitoring from K7.

Signed-off-by: Jaswinder Singh Rajput <[email protected]>
---
arch/x86/kernel/cpu/perf_counter.c | 4 ++++
1 files changed, 4 insertions(+), 0 deletions(-)

diff --git a/arch/x86/kernel/cpu/perf_counter.c b/arch/x86/kernel/cpu/perf_counter.c
index 57ae1be..719a7d5 100644
--- a/arch/x86/kernel/cpu/perf_counter.c
+++ b/arch/x86/kernel/cpu/perf_counter.c
@@ -1452,6 +1452,10 @@ static int intel_pmu_init(void)

static int amd_pmu_init(void)
{
+ /* performance-monitoring supported from K7 */
+ if (boot_cpu_data.x86 < 6)
+ return -ENODEV;
+
x86_pmu = amd_pmu;

switch (boot_cpu_data.x86) {
--
1.6.3.2


> What the heck is going on here?
>

hmm, I am really surprise how I wrote Signed-off-by line correct, I was
really very upset at that time, and you know the reason, Please help
me ;-)

Thanks,
--
JSR