[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <5947c093-cff9-f70e-af20-75bc053edf5f@linux.ibm.com>
Date: Fri, 8 Oct 2021 16:56:55 +0530
From: kajoljain <kjain@...ux.ibm.com>
To: James Clark <james.clark@....com>, acme@...nel.org,
john.garry@...wei.com, ak@...ux.intel.com,
linux-perf-users@...r.kernel.org
Cc: Nick.Forrington@....com, Andrew.Kilroy@....com,
Will Deacon <will@...nel.org>,
Mathieu Poirier <mathieu.poirier@...aro.org>,
Leo Yan <leo.yan@...aro.org>,
Mark Rutland <mark.rutland@....com>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Jiri Olsa <jolsa@...hat.com>,
Namhyung Kim <namhyung@...nel.org>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/3] perf tools: Enable strict JSON parsing
On 10/8/21 3:32 PM, James Clark wrote:
>
>
> On 08/10/2021 08:43, kajoljain wrote:
>>
>>
>> On 10/7/21 4:35 PM, James Clark wrote:
>>> After a discussion on "[PATCH 1/4] perf vendor events: Syntax corrections in Neoverse N1 json",
>>> John Garry suggested that we can just modify the parser to make it more strict. Hopefully this will
>>> remove the need to apply any future JSON comma fixup commits.
>>>
>>> Applies on top of "[PATCH v2 1/3] perf vendor events: Syntax corrections in Neoverse N1 json" on
>>> perf/core.
>>>
>>> Also available at:
>>> git clone --branch james-json-parse-fix git@....gitlab.arm.com:linux-arm/linux-jc.git
>>
>> Hi James,
>> Do we have any dependency patches on top of this patch series. I am
>> reviewing and testing it, but in both powerpc and x86 system I am
>> getting build issue. Not sure if I am missing something>
>> I am trying your changes on top of upstream perf.
>>
>> pmu-events/arch/test/test_soc/sys/uncore.json: json error Invalid
>> character inside JSON string
>
> Hi Kajol,
>
> A trailing comma was fixed in this file 3 weeks ago at b8b350a. Can you
> confirm if you have updated to get this commit on perf core?
>
> Alternately you could pull from my branch above which is up to date enough
> to include it.
Hi James,
Thanks for pointing it. Not getting build issue now.
>
> The file is in pmu-events/arch/test/ so I would expect it to fail on all platforms.
>
>> make[3]: *** [pmu-events/Build:18: pmu-events/pmu-events.c] Error 1
>> make[3]: *** Deleting file 'pmu-events/pmu-events.c'
>> make[2]: *** [Makefile.perf:667: pmu-events/pmu-events-in.o] Error 2
>> make[2]: *** Waiting for unfinished jobs....
>> make[1]: *** [Makefile.perf:238: sub-make] Error 2
>> make: *** [Makefile:70: all] Error 2
>>
>> Also, Is it possible to add line number along with file name while
>> showing this error `json error Invalid character inside JSON string`.
>> It might make it easy to fix.
>
> I can add a character number with the following fix if you think that would
> be good enough? A line number might be a bigger change and involve keeping
> track of newline characters.
Sure. I think then we can skip this change. Not sure if character
number will be helpful.
Patch-set looks good to me.
Reviewed-by Kajol Jain<kjain@...ux.ibm.com>
Thanks,
Kajol Jain
>
> diff --git a/tools/perf/pmu-events/json.c b/tools/perf/pmu-events/json.c
> index 0544398d6e2d..41a14e1543bf 100644
> --- a/tools/perf/pmu-events/json.c
> +++ b/tools/perf/pmu-events/json.c
> @@ -99,7 +99,7 @@ jsmntok_t *parse_json(const char *fn, char **map, size_t *size, int *len)
> res = jsmn_parse(&parser, *map, *size, tokens,
> sz / sizeof(jsmntok_t));
> if (res != JSMN_SUCCESS) {
> - pr_err("%s: json error %s\n", fn, jsmn_strerror(res));
> + pr_err("%s: json error at character %u '%s'\n", fn, parser.pos, jsmn_strerror(res));
> goto error_free;
> }
> if (len)
>
>
> It prints this for the same error you have above>
> pmu-events/arch/test/test_soc/sys/uncore.json: json error at character 213 'Invalid character inside JSON string'
>
> Although funnily enough after re-introducing that extra comma it doesn't fail the build for me,
> it just prints the error message. But I may have noticed some dependency tracking issues around
> the json files.
>
> James
>
Powered by blists - more mailing lists