[v5,0/4] perf: Add more syscalls to benchmark

  Tested on x86_64, arm64, mips64 and loongarch64.

Tiezhu Yang (4):
  tools x86: Keep list sorted by number in unistd_{32,64}.h
  perf bench syscall: Introduce bench_syscall_common()
  perf bench syscall: Add getpgid syscall benchmark
  perf bench syscall: Add execve syscall benchmark

 tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
 tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
 tools/perf/bench/bench.h                    |  2 +
 tools/perf/bench/syscall.c                  | 76 +++++++++++++++++++++++++++--
 tools/perf/builtin-bench.c                  |  2 +
 5 files changed, 108 insertions(+), 18 deletions(-)
  

On 11/10/2022 11:50 AM, Tiezhu Yang wrote:
> Tested on x86_64, arm64, mips64 and loongarch64.
>
> Tiezhu Yang (4):
>   tools x86: Keep list sorted by number in unistd_{32,64}.h
>   perf bench syscall: Introduce bench_syscall_common()
>   perf bench syscall: Add getpgid syscall benchmark
>   perf bench syscall: Add execve syscall benchmark
>
>  tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
>  tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
>  tools/perf/bench/bench.h                    |  2 +
>  tools/perf/bench/syscall.c                  | 76 +++++++++++++++++++++++++++--
>  tools/perf/builtin-bench.c                  |  2 +
>  5 files changed, 108 insertions(+), 18 deletions(-)
>

Ping, any more comments?

Thanks,
Tiezhu
  

On 12/03/2022 05:19 PM, Tiezhu Yang wrote:
>
>
> On 11/10/2022 11:50 AM, Tiezhu Yang wrote:
>> Tested on x86_64, arm64, mips64 and loongarch64.
>>
>> Tiezhu Yang (4):
>>   tools x86: Keep list sorted by number in unistd_{32,64}.h
>>   perf bench syscall: Introduce bench_syscall_common()
>>   perf bench syscall: Add getpgid syscall benchmark
>>   perf bench syscall: Add execve syscall benchmark
>>
>>  tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
>>  tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
>>  tools/perf/bench/bench.h                    |  2 +
>>  tools/perf/bench/syscall.c                  | 76
>> +++++++++++++++++++++++++++--
>>  tools/perf/builtin-bench.c                  |  2 +
>>  5 files changed, 108 insertions(+), 18 deletions(-)
>>
>
> Ping, any more comments?
>
> Thanks,
> Tiezhu

Hi all,

If this patch series has no value and is not acceptable,
or what should I do to update, please let me know.

Thanks,
Tiezhu
  

Hello,

On Thu, Jan 5, 2023 at 5:23 PM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>
>
>
> On 12/03/2022 05:19 PM, Tiezhu Yang wrote:
> >
> >
> > On 11/10/2022 11:50 AM, Tiezhu Yang wrote:
> >> Tested on x86_64, arm64, mips64 and loongarch64.
> >>
> >> Tiezhu Yang (4):
> >>   tools x86: Keep list sorted by number in unistd_{32,64}.h
> >>   perf bench syscall: Introduce bench_syscall_common()
> >>   perf bench syscall: Add getpgid syscall benchmark
> >>   perf bench syscall: Add execve syscall benchmark
> >>
> >>  tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
> >>  tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
> >>  tools/perf/bench/bench.h                    |  2 +
> >>  tools/perf/bench/syscall.c                  | 76
> >> +++++++++++++++++++++++++++--
> >>  tools/perf/builtin-bench.c                  |  2 +
> >>  5 files changed, 108 insertions(+), 18 deletions(-)
> >>
> >
> > Ping, any more comments?
> >
> > Thanks,
> > Tiezhu
>
> Hi all,
>
> If this patch series has no value and is not acceptable,
> or what should I do to update, please let me know.

I'm so sorry about being late a lot.
I don't have any objection to this series.

For the execve bench, I think it's actually fork + execve
then maybe it makes sense to have a fork only bench
to compare the execve part precisely.

But it can be added later, so

Acked-by: Namhyung Kim <namhyung@kernel.org>

Thanks,
Namhyung
  

On 01/07/2023 05:04 AM, Namhyung Kim wrote:
> Hello,
>
> On Thu, Jan 5, 2023 at 5:23 PM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
>>
>>
>>
>> On 12/03/2022 05:19 PM, Tiezhu Yang wrote:
>>>
>>>
>>> On 11/10/2022 11:50 AM, Tiezhu Yang wrote:
>>>> Tested on x86_64, arm64, mips64 and loongarch64.
>>>>
>>>> Tiezhu Yang (4):
>>>>   tools x86: Keep list sorted by number in unistd_{32,64}.h
>>>>   perf bench syscall: Introduce bench_syscall_common()
>>>>   perf bench syscall: Add getpgid syscall benchmark
>>>>   perf bench syscall: Add execve syscall benchmark
>>>>
>>>>  tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
>>>>  tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
>>>>  tools/perf/bench/bench.h                    |  2 +
>>>>  tools/perf/bench/syscall.c                  | 76
>>>> +++++++++++++++++++++++++++--
>>>>  tools/perf/builtin-bench.c                  |  2 +
>>>>  5 files changed, 108 insertions(+), 18 deletions(-)
>>>>
>>>
>>> Ping, any more comments?
>>>
>>> Thanks,
>>> Tiezhu
>>
>> Hi all,
>>
>> If this patch series has no value and is not acceptable,
>> or what should I do to update, please let me know.
>
> I'm so sorry about being late a lot.
> I don't have any objection to this series.
>
> For the execve bench, I think it's actually fork + execve
> then maybe it makes sense to have a fork only bench
> to compare the execve part precisely.
>
> But it can be added later, so
>
> Acked-by: Namhyung Kim <namhyung@kernel.org>
>
> Thanks,
> Namhyung
>

Hi Namhyung,

Thank you very much. I will submit another single patch
to benchmark fork after this series is applied.

Thanks,
Tiezhu
  

Em Fri, Jan 06, 2023 at 01:04:42PM -0800, Namhyung Kim escreveu:
> Hello,
> 
> On Thu, Jan 5, 2023 at 5:23 PM Tiezhu Yang <yangtiezhu@loongson.cn> wrote:
> >
> >
> >
> > On 12/03/2022 05:19 PM, Tiezhu Yang wrote:
> > >
> > >
> > > On 11/10/2022 11:50 AM, Tiezhu Yang wrote:
> > >> Tested on x86_64, arm64, mips64 and loongarch64.
> > >>
> > >> Tiezhu Yang (4):
> > >>   tools x86: Keep list sorted by number in unistd_{32,64}.h
> > >>   perf bench syscall: Introduce bench_syscall_common()
> > >>   perf bench syscall: Add getpgid syscall benchmark
> > >>   perf bench syscall: Add execve syscall benchmark
> > >>
> > >>  tools/arch/x86/include/uapi/asm/unistd_32.h | 23 ++++++---
> > >>  tools/arch/x86/include/uapi/asm/unistd_64.h | 23 ++++++---
> > >>  tools/perf/bench/bench.h                    |  2 +
> > >>  tools/perf/bench/syscall.c                  | 76
> > >> +++++++++++++++++++++++++++--
> > >>  tools/perf/builtin-bench.c                  |  2 +
> > >>  5 files changed, 108 insertions(+), 18 deletions(-)
> > >>
> > >
> > > Ping, any more comments?
> > >
> > > Thanks,
> > > Tiezhu
> >
> > Hi all,
> >
> > If this patch series has no value and is not acceptable,
> > or what should I do to update, please let me know.
> 
> I'm so sorry about being late a lot.
> I don't have any objection to this series.
> 
> For the execve bench, I think it's actually fork + execve
> then maybe it makes sense to have a fork only bench
> to compare the execve part precisely.
> 
> But it can be added later, so
> 
> Acked-by: Namhyung Kim <namhyung@kernel.org>

Applied.

- Arnaldo