[PATCH] Adding tracing documentation in user manual

Gedare Bloom gedare at rtems.org
Thu Jun 7 13:36:53 UTC 2018


On Wed, Jun 6, 2018 at 5:24 AM, Vidushi Vashishth <reachvidu at gmail.com> wrote:
>
>
>>> I am just sending a revised version with incorporations of your
>>> suggestions. I will create a fresh version 2 patch after everyone has
>>> reviewed.
>>
>>
>> If you send revised version, then please send them as a separate patch
>> with a new version. You can mention in the commit message what you changed
>> in a particular version of the patch. For example:
>>
>> https://lists.gnu.org/archive/html/qemu-devel/2018-05/msg06269.html
>>
>> There should be a section which tells me which tools I have to install.
>>
>> The new files need a proper copyright message mentioning the copyright
>> owner.
>
>
> Okay I will make these changes.
>

Also, add the images as a separate patch first, to separate these
binary files from text. Make sure the images are appropriately
licensed, or that you create them yourself. It should be preferred
that images have CC-BY-SA license on them. You'll have to generate the
necessary text for attribution if it has BY.

>> In contrast to the other documents the stuff in user uses subdirectories
>> and separate files at a low section level. Is this practical?
>
>
> I didn't quite understand this.
>

This appears more a comment from Sebastian for the general development
community, and probably belongs in a separate email thread.

> Also what are you thoughts on my idea of accomplishing function tracing
> using CTF? Do you think its a viable approach? This is written in the
> function tracing use case below:
>
> +Requirements
> +------------
> +
> +The current tracing framework provides this functionality with
> :ref:`tracebuffering`. The output is
> +provided in the form of printing on console or saving the buffer in the
> form of a bin file. In order to
> +develop this use case using CTF we need to be able to convert either the
> bin file or console output to CTF.
> +The saved bin file must also first be transported to the host from the
> target for this purpose. On the other
> +hand console output could be written to a text file which can then be
> converted to CTF.
>

I would either generate a binary CTF directly, or write an adapter for
babeltrace to ingest the rtems trace buffer's binary format. I see no
advantage to working through the ascii console output files at all. If
one can connect a console, one can stream a binary file in raw mode.

Gedare



More information about the devel mailing list