arm: g_current_regs is only used to determine if we are in irq,

[hujun260]

Summary

Why

When a context switch occurs currently, the context before the interrupt is fetched from the global variable g_current_regs and then stored in the corresponding tcb->xcp.regs. Subsequently, the new context is assigned to g_current_regs. When returning from the interrupt context, the context is retrieved from g_current_regs and used.

In reality, we do not necessarily require this intermediate variable g_current_regs to accomplish a context switch. Eliminating it avoids multiple assignments, especially during multiple context switches within interrupts, which can lead to repeated assignments.

All we need is to work directly with the context of the currently running task. This approach enhances the speed of context switching, reduces the code base, and optimizes signal handling logic. By simply leveraging the context of the active task, we can streamline the process and improve overall efficiency.

reason:
by doing this we can reduce context switch time,
When we exit from an interrupt handler, we directly use tcb->xcp.regs
Missing Information:
Related Issues: none
NuttX Apps Impact: none

Impact

Is a new feature added? NO
Is an existing feature changed? yes
g_current_regs is only used to determine if we are in irq with other functionalities removed.
We need to use up_interrupt_context for interrupt identification instead of relying on up_current_regs for that purpose.

before
size nuttx
text data bss dec hex filename
225920 409 30925 257254 3ece6 nuttx

after
text data bss dec hex filename
225604 409 30925 256938 3ebaa nuttx

szie change -316

Testing

Build Host:

• OS: Ubuntu 20.04
• CPU: x86_64
• Compiler: GCC 9.4.0

Target:

• Arch: ARM ARM64

logs:
NO change

[lupyuen]

[Experimental Bot, please feedback here]

NuttX PR Requirements Review

While the provided summary gives a high-level idea of the change, it lacks some important details.

Here's a breakdown of what's missing and how to improve the PR:

Summary:

• Clarity on the "Why": Instead of just stating functionalities were removed, explain the benefit of removing them. Why is reducing context switch time important? What kind of performance improvement does it bring?
• Specificity on "What": Which functionalities were removed exactly? This needs to be crystal clear.
• Missing Information:
  • Related Issues: Are there any related NuttX issues this PR addresses? If so, provide links.
  • NuttX Apps Impact: Does this change impact any NuttX Apps? Be specific.

Impact:

• Incomplete: You need to address all the impact points listed in the requirements. Even if the answer is "NO," state it explicitly. Don't leave any section blank. For example:
  • Is a new feature added? NO
  • Is an existing feature changed? YES (Please describe how the removal of g_current_regs functionality impacts existing features)
  • Impact on the user: NO (or YES if users need to be aware of anything)
  • ...and so on for all the impact points.

Testing:

• Insufficient Information:
  • Build Hosts: Specify the operating system, CPU architecture, and compiler used for testing.
  • Targets: Be specific about the architecture (e.g., ARM Cortex-M4) and board configuration used.
• Missing Logs:
  • Provide actual testing logs from before and after the change. This helps reviewers understand how the code behaved before and after your changes. Include relevant logs that demonstrate the performance improvement (context switch time).

Overall:

The PR is incomplete and needs more details to meet the NuttX requirements. Provide the missing information and address all the impact points clearly. Detailed testing information, including logs, is crucial for reviewers to evaluate your changes effectively.

[hujun260]

[Experimental Bot, please feedback here]

NuttX PR Requirements Review

While the provided summary gives a high-level idea of the change, it lacks some important details.

Here's a breakdown of what's missing and how to improve the PR:

Summary:

• Clarity on the "Why": Instead of just stating functionalities were removed, explain the benefit of removing them. Why is reducing context switch time important? What kind of performance improvement does it bring?

• Specificity on "What": Which functionalities were removed exactly? This needs to be crystal clear.

• Missing Information:

  • Related Issues: Are there any related NuttX issues this PR addresses? If so, provide links.
  • NuttX Apps Impact: Does this change impact any NuttX Apps? Be specific.

Impact:

• Incomplete: You need to address all the impact points listed in the requirements. Even if the answer is "NO," state it explicitly. Don't leave any section blank. For example:

  • Is a new feature added? NO
  • Is an existing feature changed? YES (Please describe how the removal of g_current_regs functionality impacts existing features)
  • Impact on the user: NO (or YES if users need to be aware of anything)
  • ...and so on for all the impact points.

Testing:

• Insufficient Information:

  • Build Hosts: Specify the operating system, CPU architecture, and compiler used for testing.
  • Targets: Be specific about the architecture (e.g., ARM Cortex-M4) and board configuration used.

• Missing Logs:

  • Provide actual testing logs from before and after the change. This helps reviewers understand how the code behaved before and after your changes. Include relevant logs that demonstrate the performance improvement (context switch time).

Overall:

The PR is incomplete and needs more details to meet the NuttX requirements. Provide the missing information and address all the impact points clearly. Detailed testing information, including logs, is crucial for reviewers to evaluate your changes effectively.

i update the pr

[xiaoxiang781216]

please fix conflict, @hujun260

[github-actions]

This PR exceeds the recommended size of 1000 lines. Please make sure you are NOT addressing multiple issues with one PR. Note this PR might be rejected due to its size.

[github-actions]

This PR exceeds the recommended size of 1000 lines. Please make sure you are NOT addressing multiple issues with one PR. Note this PR might be rejected due to its size.

[masayuki2009]

@hujun260
I noticed that spresense:smp and raspberrypi-pico:smp do not boot with this PR.

[hujun260]

@hujun260 I noticed that spresense:smp and raspberrypi-pico:smp do not boot with this PR.

ok i'll look into this issue

[hujun260]

#13598 fix this issue