sched_processtimer: use small lock to protect g_timer_interval and g_timer_tick

[hujun260]

Summary

sched_processtimer: use small lock to protect g_timer_interval and g_timer_tick

reason:
We would like to replace the critical section with a small lock.

Impact

sched_processtimer

Testing

ci ostest

[nuttxpr]

[Experimental Bot, please feedback here]

No, this PR does not fully meet the NuttX requirements. While it provides a summary of the what, it lacks sufficient detail on the why and how. The Impact and Testing sections are also severely lacking.

Here's a breakdown of what's missing:

• Summary:

  • Missing "Why": Why is replacing the critical section with a small lock necessary? What problem does it solve? Performance improvement? Deadlock avoidance? Explain the rationale.
  • Missing "How": What specific changes were made to the code to implement the small lock? Which lock is used? How does this interact with other locking mechanisms?

• Impact:

  • Insufficient Detail: Simply stating "sched_processtimer" is not enough. Address all the impact points. At minimum, explain if this impacts users, the build, hardware, documentation, security, or compatibility. Even if the answer is "NO" for each, explicitly state it. For example: "Impact on user: NO".

• Testing:

  • Insufficient Detail: "ci ostest" is not sufficient. Provide specifics:
    • Build Host(s): What operating system, CPU architecture, and compiler version were used for building?
    • Target(s): What architecture and board/configuration were used for testing? Was this tested on a simulator or real hardware?
    • Logs: The provided log sections are empty. Include actual logs demonstrating the behavior before and after the change. Show how the change addresses the problem described in the "Why" section of the summary.

In short, the PR needs significantly more detail to be considered complete. It needs to clearly explain the motivation, implementation, and testing of the change.

[anchao]

nxsched_timer_start() will access the global variables in lower harlf driver, using local spinlock cannot guarantee the same effect as csection.

[hujun260]

nxsched_timer_start will call up_timer_start or up_alarm_start
I believe that these two functions should be responsible for adding their own critical sections. In fact, this is probably how it's done in most cases.

However, there may be some overlooked scenarios. Currently, we can add an additional critical section in this function. Once all the timer driver codes have been replaced with small locks in the future, this critical section can be removed.

[hujun260]

[xiaoxiang781216]

why mix the critical section and spinlock?

[hujun260]

Perhaps we need to protect up_timer_tick_start.