[PATCH] perf: make perf script actually spit out backtraces for events

[PATCH] perf: make perf script actually spit out backtraces for events

[Josef Bacik]

We are trying to debug a problem internally and noticed that if we use perf
record -g -e <tracepoint> and then try to use perf script to get the events and
their backtraces that we weren't getting the backtraces, even though they were
collected and would show up with perf report.  For some reason the tracepoint
print stuff didn't have the right things set in order to print out backtraces.
I have no idea why this is as adding them in there makes the backtraces come out
properly, and if we don't use the -g option perf script still acts properly with
those options set.  Hopefully this is the correct fix, it has fixed the problem
for us.  Thanks,

Signed-off-by: Josef Bacik <jbacik@fb.com>
---
NOTE: I looked at MAINTAINERS to see if there was a perf list and it didn't look
like there was one, sorry if this is the wrong place to send patches to.

 tools/perf/builtin-script.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index baf1798..6d14c06 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -99,7 +99,9 @@ static struct {
 
 		.fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID |
 				  PERF_OUTPUT_CPU | PERF_OUTPUT_TIME |
-				  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE,
+				  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE |
+				  PERF_OUTPUT_IP | PERF_OUTPUT_SYM |
+				  PERF_OUTPUT_DSO,
 	},
 
 	[PERF_TYPE_RAW] = {
-- 
1.8.3.1

Re: [PATCH] perf: make perf script actually spit out backtraces for events

[David Ahern]

On 3/26/14, 12:57 PM, Josef Bacik wrote:
> We are trying to debug a problem internally and noticed that if we use perf
> record -g -e <tracepoint> and then try to use perf script to get the events and
> their backtraces that we weren't getting the backtraces, even though they were
> collected and would show up with perf report.  For some reason the tracepoint
> print stuff didn't have the right things set in order to print out backtraces.
> I have no idea why this is as adding them in there makes the backtraces come out
> properly, and if we don't use the -g option perf script still acts properly with
> those options set.  Hopefully this is the correct fix, it has fixed the problem
> for us.  Thanks,

You can manually request fields using the -f option.

That said ...

>
> Signed-off-by: Josef Bacik <jbacik@fb.com>
> ---
> NOTE: I looked at MAINTAINERS to see if there was a perf list and it didn't look
> like there was one, sorry if this is the wrong place to send patches to.
>
>   tools/perf/builtin-script.c | 4 +++-
>   1 file changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
> index baf1798..6d14c06 100644
> --- a/tools/perf/builtin-script.c
> +++ b/tools/perf/builtin-script.c
> @@ -99,7 +99,9 @@ static struct {
>
>   		.fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID |
>   				  PERF_OUTPUT_CPU | PERF_OUTPUT_TIME |
> -				  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE,
> +				  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE |
> +				  PERF_OUTPUT_IP | PERF_OUTPUT_SYM |
> +				  PERF_OUTPUT_DSO,

A change has been applied to upstream:
https://lkml.org/lkml/2013/11/30/63


David

Re: [PATCH] perf: make perf script actually spit out backtraces for events

[Josef Bacik]

On 03/26/2014 03:27 PM, David Ahern wrote:
> On 3/26/14, 12:57 PM, Josef Bacik wrote:
>> We are trying to debug a problem internally and noticed that if we use
>> perf
>> record -g -e <tracepoint> and then try to use perf script to get the
>> events and
>> their backtraces that we weren't getting the backtraces, even though
>> they were
>> collected and would show up with perf report.  For some reason the
>> tracepoint
>> print stuff didn't have the right things set in order to print out
>> backtraces.
>> I have no idea why this is as adding them in there makes the
>> backtraces come out
>> properly, and if we don't use the -g option perf script still acts
>> properly with
>> those options set.  Hopefully this is the correct fix, it has fixed
>> the problem
>> for us.  Thanks,
>
> You can manually request fields using the -f option.
>
> That said ...
>
>>
>> Signed-off-by: Josef Bacik <jbacik@fb.com>
>> ---
>> NOTE: I looked at MAINTAINERS to see if there was a perf list and it
>> didn't look
>> like there was one, sorry if this is the wrong place to send patches to.
>>
>>   tools/perf/builtin-script.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
>> index baf1798..6d14c06 100644
>> --- a/tools/perf/builtin-script.c
>> +++ b/tools/perf/builtin-script.c
>> @@ -99,7 +99,9 @@ static struct {
>>
>>           .fields = PERF_OUTPUT_COMM | PERF_OUTPUT_TID |
>>                     PERF_OUTPUT_CPU | PERF_OUTPUT_TIME |
>> -                  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE,
>> +                  PERF_OUTPUT_EVNAME | PERF_OUTPUT_TRACE |
>> +                  PERF_OUTPUT_IP | PERF_OUTPUT_SYM |
>> +                  PERF_OUTPUT_DSO,
>
> A change has been applied to upstream:
> https://urldefense.proofpoint.com/v1/url?u=https://lkml.org/lkml/2013/11/30/63&k=ZVNjlDMF0FElm4dQtryO4A%3D%3D%0A&r=cKCbChRKsMpTX8ybrSkonQ%3D%3D%0A&m=kB8XhiqkBzGE%2BTCnMhkJQP%2BjerKZ5bWPi%2BqGVVwGKPg%3D%0A&s=c232c9e9ea4eb35f56ba43dd8e7e70589e016430a7b4a5a89f7c938d05bd7b7a
>

Argh sorry I was testing on my tree but I'm still 3.13 based.  Thanks,

Josef