2022-03-21 22:06:07

by Steven Rostedt

[permalink] [raw]
Subject: [GIT PULL] tracing/rtla: Updates to the RTLA tool


Linus,

I'm sending this as a separate pull request from my normal tracing tree, as
it had dependencies to commits in your tree but not in my development
branch. As it only affects the tools/ directory, I kept it separate as it
has no changes to the kernel itself.

Real Time Linux Analysis Tool updates for 5.18

Changes to RTLA:

- Support for adjusting tracing_threashold

- Add -a (auto) option to make it easier for users to debug in the field

- Add -e option to add more events to the trace

- Add --trigger option to add triggers to events

- Add --filter option to filter events

- Add support to save histograms to the file

- Add --dma-latency to set /dev/cpu_dma_latency

- Other fixes and cleanups

Please pull the latest trace-rtla-v5.18 tree, which can be found at:


git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git
trace-rtla-v5.18

Tag SHA1: cc921ea64d6df1bb8955ff5762dbcdc408e34264
Head SHA1: 75016ca3acd0de79868ef5b0694195fe05288ade


Daniel Bristot de Oliveira (15):
rtla/osnoise: Add support to adjust the tracing_thresh
rtla/osnoise: Add an option to set the threshold
rtla/osnoise: Add the automatic trace option
rtla/timerlat: Add the automatic trace option
rtla/trace: Add trace events helpers
rtla: Add -e/--event support
rtla/trace: Add trace event trigger helpers
rtla: Add --trigger support
rtla/trace: Add trace event filter helpers
rtla: Add --filter support
rtla/trace: Save event histogram output to a file
rtla: Check for trace off also in the trace instance
rtla/osnoise: Fix osnoise hist stop tracing message
rtla/timerlat: Add --dma-latency option
rtla: Tools main loop cleanup

----
Documentation/tools/rtla/common_options.rst | 19 ++
.../tools/rtla/common_osnoise_options.rst | 10 +
.../tools/rtla/common_timerlat_options.rst | 12 +
tools/tracing/rtla/src/osnoise.c | 83 +++++
tools/tracing/rtla/src/osnoise.h | 8 +
tools/tracing/rtla/src/osnoise_hist.c | 95 +++++-
tools/tracing/rtla/src/osnoise_top.c | 91 +++++-
tools/tracing/rtla/src/timerlat_hist.c | 108 ++++++-
tools/tracing/rtla/src/timerlat_top.c | 103 +++++-
tools/tracing/rtla/src/trace.c | 345 +++++++++++++++++++++
tools/tracing/rtla/src/trace.h | 23 ++
tools/tracing/rtla/src/utils.c | 33 ++
tools/tracing/rtla/src/utils.h | 1 +
13 files changed, 903 insertions(+), 28 deletions(-)
---------------------------


Subject: Re: [GIT PULL] tracing/rtla: Updates to the RTLA tool

On 3/23/22 20:43, Steven Rostedt wrote:
>> That said, mentioning the perf tools, I wish the tracing tools would
>> do a bit more package checking and helpful error messages too, rather
>> than just fail with
>>
>> fatal error: tracefs.h: No such file or directory
>>
>> when (presumably) libtracefs and friends are needed.
> Yes exactly.
>
> Daniel, care to update the Makefile to check for libtracefs and friends and
> give a more informative message when it fails?
>
> You can look at the trace-cmd Makefile that does so.

Working on that!

-- Daniel

2022-03-24 14:18:38

by Steven Rostedt

[permalink] [raw]
Subject: Re: [GIT PULL] tracing/rtla: Updates to the RTLA tool

On Wed, 23 Mar 2022 11:15:30 -0700
Linus Torvalds <[email protected]> wrote:

> That said, mentioning the perf tools, I wish the tracing tools would
> do a bit more package checking and helpful error messages too, rather
> than just fail with
>
> fatal error: tracefs.h: No such file or directory
>
> when (presumably) libtracefs and friends are needed.

Yes exactly.

Daniel, care to update the Makefile to check for libtracefs and friends and
give a more informative message when it fails?

You can look at the trace-cmd Makefile that does so.

-- Steve

2022-03-24 15:42:26

by pr-tracker-bot

[permalink] [raw]
Subject: Re: [GIT PULL] tracing/rtla: Updates to the RTLA tool

The pull request you sent on Mon, 21 Mar 2022 10:30:35 -0400:

> git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace.git trace-rtla-v5.18

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/20f463fb38686dd3fe7e6903cab56bdbbf756238

Thank you!

--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html

2022-03-25 06:19:40

by Linus Torvalds

[permalink] [raw]
Subject: Re: [GIT PULL] tracing/rtla: Updates to the RTLA tool

On Mon, Mar 21, 2022 at 7:30 AM Steven Rostedt <[email protected]> wrote:
>
> I'm sending this as a separate pull request from my normal tracing tree, as
> it had dependencies to commits in your tree but not in my development
> branch. As it only affects the tools/ directory, I kept it separate as it
> has no changes to the kernel itself.

Yes, please keep the tooling updates separate from the kernel ones,
the way we now for some time have for the perf tooling too.

That said, mentioning the perf tools, I wish the tracing tools would
do a bit more package checking and helpful error messages too, rather
than just fail with

fatal error: tracefs.h: No such file or directory

when (presumably) libtracefs and friends are needed.

Hmm?

Linus

Subject: Re: [GIT PULL] tracing/rtla: Updates to the RTLA tool

On 3/23/22 19:15, Linus Torvalds wrote:
> On Mon, Mar 21, 2022 at 7:30 AM Steven Rostedt <[email protected]> wrote:
>>
>> I'm sending this as a separate pull request from my normal tracing tree, as
>> it had dependencies to commits in your tree but not in my development
>> branch. As it only affects the tools/ directory, I kept it separate as it
>> has no changes to the kernel itself.
>
> Yes, please keep the tooling updates separate from the kernel ones,
> the way we now for some time have for the perf tooling too.
>
> That said, mentioning the perf tools, I wish the tracing tools would
> do a bit more package checking and helpful error messages too, rather
> than just fail with
>
> fatal error: tracefs.h: No such file or directory
>
> when (presumably) libtracefs and friends are needed.
>
> Hmm?

Yes, we need to improve the Makefile.

It depends on procps-ng-devel and python3-docutils packages and the latest
libtracefs (1.3) and libtraceevent (1.5). There is info about it in the README,
but we definitely need to improve the Makefile... I will prioritize that.

-- Daniel
> Linus