Re: [PATCH] perf vendor events arm64 AmpereOneX: Fix misspelled PublicDescription keys
From: James Clark
Date: Wed Nov 11 2026 - 09:03:23 EST
On 11/11/2026 13:35, TAKANO Mitsuhiro wrote:
On 2026-11-11 10:16, James Clark wrote:
IMO I would rather have something that's incredibly strict and simple,
or nothing at all. Having to work around edge cases and report near
misses doesn't sound great.
Agreed, and the near-miss report is gone. The series is posted again:
https://lore.kernel.org/linux-perf-users/20261111132622.2349165-1-takano32@xxxxxxxxx/
The check is now a fixed list of the 51 keys a json object under arch/ may
carry: 47 of them jevents.py reads, one more metric.py reads, and three are
read by nothing although the tables hold them (Counter, CountersNumFixed,
CountersNumGeneric). A key outside that list is reported with the object
that carries it. There is no cut-off, no similarity and nothing that
depends on what a run of the generator happened to read, so there is no
edge case left to work around.
The linked patches seem like a bigger change than what I'd expect. I'm not sure that's necessary for something conceptually as simple as "validate the json files".
I'm also a bit lost about the relationship between "[PATCH 0/3] perf jevents: Report the JSON keys that jevents.py never reads" and "[PATCH] perf jevents: Check the JSON keys that jevents.py never reads".
It still runs as "make check-jevents" and as a perf test rather than as part
of the build, and on the tree today it reports exactly the two misspellings
we have, the 16 in this file and 3 in the AMD tables, so it wants those
fixes in front of it or together with it.
The one thing I did not do is put the list in a json schema file, because
perf's build would then require jsonschema, which it does not use today and
which is not in every builder's python. The list holds the same names that
It could run if it exists and skip if not. Or we could make it a requirement, but that would require some agreement. The docs builder already has a pip requirements.txt file.
Even without jsonschema I can't see why a validator can't be written in less than 10 lines.
a schema with "additionalProperties": false would say. If you would rather
have the schema file and take the dependency, it is a small change from
here.