trace: replace all *_wo_note to *_notrace to make api define more approachable

[anchao]

Summary

trace: replace all *_wo_note to *_notrace to make api define more approachable

Follow the linux naming style:

https://github.com/torvalds/linux/blob/master/include/linux/preempt.h#L237-L242
https://github.com/torvalds/linux/blob/master/include/linux/srcu.h#L221

Signed-off-by: chao an [email protected]

Impact

N/A

Testing

ci-check

[nuttxpr]

[Experimental Bot, please feedback here]

No, this PR does not fully meet the NuttX requirements. While it provides a summary of the change and testing information, it lacks crucial details.

Here's a breakdown of what's missing:

• Insufficient Summary: While the summary explains what was changed, it doesn't fully explain why. Why is the Linux naming style preferred? What problem does the current naming create? What benefit does the new naming provide? This is critical for reviewers to understand the motivation behind the change.

• Impact Section is too brief: "N/A" is not acceptable. Even if there's minimal impact, each item should be addressed with a "NO" and a brief explanation justifying why there is no impact. For example, "Impact on user: NO - This is a purely internal naming change with no user-facing effects." This demonstrates that the impact has been considered. If the PR genuinely has no impact on any of these elements, that is rather uncommon. Changes to the naming of APIs are likely to have implications on documentation or backward compatibility as a bare minimum.

• Missing Testing Details: "ci-check" is insufficient. While CI is important, the PR needs to specify what was tested and how. Provide specific test cases or commands used. "Before" and "After" logs are required, even if they are identical (to demonstrate no regression). If the CI covers the testing completely, then link to the specific CI run and highlight the relevant parts that demonstrate the change works. Also include details on the build host and target used for testing.

In short, the PR needs to be more thorough and explicit in addressing all the requirements, even if the impact or changes are minimal. The goal is to provide enough information for reviewers to quickly understand and assess the change.

[xiaoxiang781216]

but nuttx trace framework name note not trace.

[anchao]

But higher-level interfaces are trace models:
https://github.com/apache/nuttx/blob/master/include/nuttx/trace.h
could we rename trace framework from note to trace?

[xiaoxiang781216]

either way is fine, but we need be consistent.

[anchao]

I think this may require 2 stages to finish this work:

1. refine current code base, rename caller interface from note to trace.
2. rename note driver module to trace

[xiaoxiang781216]

yes, I am fine with the proposal. please update the document too.

[xiaoxiang781216]

the patch doesn't fix the problem I rise.

[anchao]

wo_ was added by Xiaomi. It is not in the history of nuttx. Some developers mistakenly believe that wo is the abbreviation of write only. The abbreviation wo(without) is not an English convention. I think we don't need to make any document changes.

[xiaoxiang781216]

Sorry, it isn't added by us but just follow the original code base. The first patch comes from:

commit ab3fa890239399de51e5482e595a751fee249b1c
Author: Masayuki Ishikawa <[email protected]>
Date:   Wed Jan 17 13:08:03 2018 +0900

    SMP: Introduce spin_lock_wo_note() and spin_unlock_wo_note()
    
    These APIs are used in sched_note.c to protect instumentation data.
    The deffrence between these APIs to exsiting spin_lock() and spin_unlock()
    is that they do not perform insturumentation to avoid recursive call
    when SCHED_INSTRUMENTATION_SPINLOCKS=y.

but xxnote is more consistent then xxtrace, since nuttx name the profile api as sched_note not trace. If you want to change the term, please change all places.

[anchao]

But why would you allow trace.h to be merged into nuttx? And the kernel code is also using similar APIs?

https://github.com/apache/nuttx/blob/master/include/nuttx/trace.h
https://github.com/apache/nuttx/blob/master/sched/task/task_init.c#L97

should we rename all trace to note?

[xiaoxiang781216]

Yes, @Gary-Hobson please fix the inconsistence.