Re: [PATCH v6 4/4] perf vendor events arm64: Add ThunderX2 implementation defined pmu core events

[Ganapatrao Kulkarni]

On Wed, Oct 11, 2017 at 7:25 PM, Will Deacon <will.deacon@xxxxxxx> wrote:
> On Wed, Oct 11, 2017 at 03:24:31PM +0200, Robert Richter wrote:
>> On 11.10.17 13:19:27, Will Deacon wrote:
>> > On Tue, Aug 29, 2017 at 02:47:30PM +0200, Robert Richter wrote:
>> > > Shaokun,
>> > >
>> > > On 29.08.17 17:26:00, Zhangshaokun wrote:
>> > > > On 2017/8/24 20:03, Ganapatrao Kulkarni wrote:
>> > > > > This is not a full event list, but a short list of useful events.
>> > > > >
>> > > > > Signed-off-by: Ganapatrao Kulkarni <ganapatrao.kulkarni@xxxxxxxxxx>
>> > > > > ---
>> > > > >  tools/perf/pmu-events/arch/arm64/mapfile.csv       | 15 ++++++
>> > > > >  .../arm64/thunderx2/implementation-defined.json    | 62 ++++++++++++++++++++++
>> > > > >  2 files changed, 77 insertions(+)
>> > > > >  create mode 100644 tools/perf/pmu-events/arch/arm64/mapfile.csv
>> > > > >  create mode 100644 tools/perf/pmu-events/arch/arm64/thunderx2/implementation-defined.json
>> > > > >
>> > > >
>> > > > I saw you also used thunderx2 in tools/perf/pmu-events/arch/arm64/, how about John's suggestion
>> > > > that would use vendor sub-folder?
>> > > > Of course, appreciate maintainer's comments.
>> > >
>> > > this would just add another level of subdirectories. I rather would
>> > > prefer to have a per platform dir comparable to what is listed in
>> > >
>> > >  arch/arm64/Kconfig.platforms
>> > >
>> > > This is the same as Ganapat has implemented it.
>> >
>> > FWIW, I agree with Zhangshaokun here that a silicon vendor subdirectory
>> > would organise things better. It also matches what we do for
>> > arch/arm64/boot/dts/
>>
>> A file structure like:
>>
>>  tools/perf/pmu-events/arch/arm64/cavium/thunderx2/implementation-defined.json
>>
>> looks quite horible. In contrast to dts dir we will need another
>> subdir for each platform and will then have two levels of subdirs.
>
> In which case, call the file thunderx2-imp-def.json or something. The
> advantage then is that the SoC names are namespaced by vendor, which is
> the best way to avoid confusion imo.

thanks, i can rename json file as suggested.

>
> Will

thanks
Ganapat