[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d7056fb8-d218-4ef4-ec45-a9af8e13045c@huawei.com>
Date: Thu, 12 May 2022 16:32:09 +0100
From: John Garry <john.garry@...wei.com>
To: Nick Forrington <nick.forrington@....com>,
<linux-kernel@...r.kernel.org>, <linux-perf-users@...r.kernel.org>,
<acme@...nel.org>
CC: Will Deacon <will@...nel.org>,
Mathieu Poirier <mathieu.poirier@...aro.org>,
Leo Yan <leo.yan@...aro.org>,
Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...nel.org>,
"Namhyung Kim" <namhyung@...nel.org>,
Andi Kleen <ak@...ux.intel.com>,
Kajol Jain <kjain@...ux.ibm.com>,
James Clark <james.clark@....com>,
Andrew Kilroy <andrew.kilroy@....com>
Subject: Re: [PATCH 01/20] perf vendors events arm64: Arm Cortex-A5
On 10/05/2022 11:47, Nick Forrington wrote:
> --- a/tools/perf/pmu-events/arch/arm64/common-and-microarch.json
> +++ b/tools/perf/pmu-events/arch/arm64/common-and-microarch.json
> @@ -35,6 +35,18 @@
> "EventName": "L1D_TLB_REFILL",
> "BriefDescription": "Attributable Level 1 data TLB refill"
> },
This is a comment on the general situation of pmu-events support for arm64:
> + {
> + "PublicDescription": "Instruction architecturally executed, condition code check pass, load",
> + "EventCode": "0x06",
> + "EventName": "LD_RETIRED",
> + "BriefDescription": "Instruction architecturally executed, condition code check pass, load"
> + },
> + {
> +[
> + {
> + "ArchStdEvent": "L1I_CACHE_REFILL"
> + },
The JSONs for some cores list these common arch events and some don't.
The effect (if we do) is that the perf tool creates the alias for the
event and we get all the event info in "perf list", which is useful.
It would be good to have consistency here, but so many arm
implementations exist and it's tricky to have all cores supported in
pmu-events. So I had a patch series which makes perf read the armv8 pmu
sysfs event file to learn all the events which the core supports and
create the aliases from that. So, in this, we don't require the JSONs to
list these events explicitly. Maybe I'll revisit it soon.
Thanks,
John
Powered by blists - more mailing lists