* [PATCH] perf data convert: Fix segfault when converting to json on arm64
@ 2024-01-11 23:29 Ilkka Koskinen
2024-01-12 11:35 ` James Clark
0 siblings, 1 reply; 4+ messages in thread
From: Ilkka Koskinen @ 2024-01-11 23:29 UTC (permalink / raw)
To: Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Namhyung Kim,
Ian Rogers, Adrian Hunter
Cc: Ilkka Koskinen, linux-perf-users, linux-kernel
Arm64 doesn't have Model in /proc/cpuinfo and, thus, cpu_desc doesn't get
assigned.
Running
$ perf data convert --to-json perf.data.json
ends up calling output_json_string() with NULL pointer, which causes a
segmentation fault.
Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
---
tools/perf/util/data-convert-json.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
index 5bb3c2ba95ca..5d6de1cef546 100644
--- a/tools/perf/util/data-convert-json.c
+++ b/tools/perf/util/data-convert-json.c
@@ -97,6 +97,11 @@ static void output_json_format(FILE *out, bool comma, int depth, const char *for
static void output_json_key_string(FILE *out, bool comma, int depth,
const char *key, const char *value)
{
+ if (!value) {
+ pr_info("No value set for key %s\n", key);
+ return;
+ }
+
output_json_delimiters(out, comma, depth);
output_json_string(out, key);
fputs(": ", out);
--
2.43.0
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] perf data convert: Fix segfault when converting to json on arm64
2024-01-11 23:29 [PATCH] perf data convert: Fix segfault when converting to json on arm64 Ilkka Koskinen
@ 2024-01-12 11:35 ` James Clark
2024-01-16 21:53 ` Namhyung Kim
0 siblings, 1 reply; 4+ messages in thread
From: James Clark @ 2024-01-12 11:35 UTC (permalink / raw)
To: Ilkka Koskinen
Cc: linux-perf-users, linux-kernel, Peter Zijlstra, Ingo Molnar,
Arnaldo Carvalho de Melo, Mark Rutland, Alexander Shishkin,
Jiri Olsa, Namhyung Kim, Ian Rogers, Adrian Hunter
On 11/01/2024 23:29, Ilkka Koskinen wrote:
> Arm64 doesn't have Model in /proc/cpuinfo and, thus, cpu_desc doesn't get
> assigned.
>
> Running
> $ perf data convert --to-json perf.data.json
>
> ends up calling output_json_string() with NULL pointer, which causes a
> segmentation fault.
>
> Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
> ---
> tools/perf/util/data-convert-json.c | 5 +++++
> 1 file changed, 5 insertions(+)
>
> diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
> index 5bb3c2ba95ca..5d6de1cef546 100644
> --- a/tools/perf/util/data-convert-json.c
> +++ b/tools/perf/util/data-convert-json.c
> @@ -97,6 +97,11 @@ static void output_json_format(FILE *out, bool comma, int depth, const char *for
> static void output_json_key_string(FILE *out, bool comma, int depth,
> const char *key, const char *value)
> {
> + if (!value) {
> + pr_info("No value set for key %s\n", key);
> + return;
> + }
> +
> output_json_delimiters(out, comma, depth);
> output_json_string(out, key);
> fputs(": ", out);
It looks like this would hide new errors on any of the other fields that
output_json_key_string() is called on. Maybe it would be better to only
wrap the call to output cpu_desc with the if? If that's the only one
that we think is optional, and even better only do it for arm64.
I mention this because the test for 'perf data convert' only checks for
valid json syntax, but not any fields. So we might want to avoid others
going missing.
Thanks
James
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] perf data convert: Fix segfault when converting to json on arm64
2024-01-12 11:35 ` James Clark
@ 2024-01-16 21:53 ` Namhyung Kim
2024-01-17 4:47 ` Ilkka Koskinen
0 siblings, 1 reply; 4+ messages in thread
From: Namhyung Kim @ 2024-01-16 21:53 UTC (permalink / raw)
To: Ilkka Koskinen
Cc: James Clark, linux-perf-users, linux-kernel, Peter Zijlstra,
Ingo Molnar, Arnaldo Carvalho de Melo, Mark Rutland,
Alexander Shishkin, Jiri Olsa, Ian Rogers, Adrian Hunter
Hello,
On Fri, Jan 12, 2024 at 3:35 AM James Clark <james.clark@arm.com> wrote:
>
>
>
> On 11/01/2024 23:29, Ilkka Koskinen wrote:
> > Arm64 doesn't have Model in /proc/cpuinfo and, thus, cpu_desc doesn't get
> > assigned.
> >
> > Running
> > $ perf data convert --to-json perf.data.json
> >
> > ends up calling output_json_string() with NULL pointer, which causes a
> > segmentation fault.
> >
> > Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
> > ---
> > tools/perf/util/data-convert-json.c | 5 +++++
> > 1 file changed, 5 insertions(+)
> >
> > diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
> > index 5bb3c2ba95ca..5d6de1cef546 100644
> > --- a/tools/perf/util/data-convert-json.c
> > +++ b/tools/perf/util/data-convert-json.c
> > @@ -97,6 +97,11 @@ static void output_json_format(FILE *out, bool comma, int depth, const char *for
> > static void output_json_key_string(FILE *out, bool comma, int depth,
> > const char *key, const char *value)
> > {
> > + if (!value) {
> > + pr_info("No value set for key %s\n", key);
> > + return;
> > + }
> > +
> > output_json_delimiters(out, comma, depth);
> > output_json_string(out, key);
> > fputs(": ", out);
>
>
> It looks like this would hide new errors on any of the other fields that
> output_json_key_string() is called on. Maybe it would be better to only
> wrap the call to output cpu_desc with the if? If that's the only one
> that we think is optional, and even better only do it for arm64.
>
> I mention this because the test for 'perf data convert' only checks for
> valid json syntax, but not any fields. So we might want to avoid others
> going missing.
Makes sense. Ilkka, can you send v2 with this?
Thanks,
Namhyung
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] perf data convert: Fix segfault when converting to json on arm64
2024-01-16 21:53 ` Namhyung Kim
@ 2024-01-17 4:47 ` Ilkka Koskinen
0 siblings, 0 replies; 4+ messages in thread
From: Ilkka Koskinen @ 2024-01-17 4:47 UTC (permalink / raw)
To: Namhyung Kim
Cc: Ilkka Koskinen, James Clark, linux-perf-users, linux-kernel,
Peter Zijlstra, Ingo Molnar, Arnaldo Carvalho de Melo,
Mark Rutland, Alexander Shishkin, Jiri Olsa, Ian Rogers,
Adrian Hunter
[-- Attachment #1: Type: text/plain, Size: 2085 bytes --]
On Tue, 16 Jan 2024, Namhyung Kim wrote:
> Hello,
>
> On Fri, Jan 12, 2024 at 3:35 AM James Clark <james.clark@arm.com> wrote:
>>
>>
>>
>> On 11/01/2024 23:29, Ilkka Koskinen wrote:
>>> Arm64 doesn't have Model in /proc/cpuinfo and, thus, cpu_desc doesn't get
>>> assigned.
>>>
>>> Running
>>> $ perf data convert --to-json perf.data.json
>>>
>>> ends up calling output_json_string() with NULL pointer, which causes a
>>> segmentation fault.
>>>
>>> Signed-off-by: Ilkka Koskinen <ilkka@os.amperecomputing.com>
>>> ---
>>> tools/perf/util/data-convert-json.c | 5 +++++
>>> 1 file changed, 5 insertions(+)
>>>
>>> diff --git a/tools/perf/util/data-convert-json.c b/tools/perf/util/data-convert-json.c
>>> index 5bb3c2ba95ca..5d6de1cef546 100644
>>> --- a/tools/perf/util/data-convert-json.c
>>> +++ b/tools/perf/util/data-convert-json.c
>>> @@ -97,6 +97,11 @@ static void output_json_format(FILE *out, bool comma, int depth, const char *for
>>> static void output_json_key_string(FILE *out, bool comma, int depth,
>>> const char *key, const char *value)
>>> {
>>> + if (!value) {
>>> + pr_info("No value set for key %s\n", key);
>>> + return;
>>> + }
>>> +
>>> output_json_delimiters(out, comma, depth);
>>> output_json_string(out, key);
>>> fputs(": ", out);
>>
>>
>> It looks like this would hide new errors on any of the other fields that
>> output_json_key_string() is called on. Maybe it would be better to only
>> wrap the call to output cpu_desc with the if? If that's the only one
>> that we think is optional, and even better only do it for arm64.
>>
>> I mention this because the test for 'perf data convert' only checks for
>> valid json syntax, but not any fields. So we might want to avoid others
>> going missing.
>
> Makes sense. Ilkka, can you send v2 with this?
I initially considered the choice James suggested but I kind of thought
that pr_info() might be enough. However, I don't have strong prefence on
either. I'll send an updated version soon.
Cheers, Ilkka
>
> Thanks,
> Namhyung
>
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2024-01-17 4:47 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2024-01-11 23:29 [PATCH] perf data convert: Fix segfault when converting to json on arm64 Ilkka Koskinen
2024-01-12 11:35 ` James Clark
2024-01-16 21:53 ` Namhyung Kim
2024-01-17 4:47 ` Ilkka Koskinen
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).