Revert "sem: change sem wait to atomic operation"

[yamt]

Summary

This reverts commit befe298.

Because a few regressions have been reported and
it likely will take some time to fix them:

• for some configurations, semaphore can be used on the special memory region, where atomic access is not available. cf. sched/semaphore: change semcount type to int #14625

• include/nuttx/lib/stdatomic.h is not compatible with the C11 semantics, which the change in question relies on. cf. semaphore: Fix a few regressions #14755

Impact

Testing

tested with esp32s3-devkit:smp ostest

[nuttxpr]

[Experimental Bot, please feedback here]

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

Here's a breakdown of what's missing:

• Summary: While it explains why the revert is happening, it lacks details on what functionality befe298 introduced in the first place. What part of the code did that commit modify? What was its intended purpose? This context is essential for reviewers.
• Impact: This section is completely empty. Reverting a change can have significant impacts. The PR needs to address all the impact points, even if the answer is "NO". Specifically, it needs to detail the impact the reverted functionality had so reviewers can assess the consequences of removing it. For example, what features are being disabled/re-enabled? Are any build changes being undone?
• Testing: While the target is mentioned, the build host information is missing. Also, "works as intended" is vague. What was the expected behavior before and after the revert? The provided logs should demonstrate the fix for the reported regressions (ideally, show the failing behavior before the revert and the working behavior after).

To make this PR compliant, the author should:

1. Expand the Summary: Explain what functionality befe298 introduced and what parts of the code it affected.
2. Complete the Impact section: Address all impact points. Consider the impact of removing the functionality added by the reverted commit.
3. Provide complete testing details: Include the build host information (OS, CPU, compiler). Clarify the expected behavior and ensure the logs demonstrate the resolution of the mentioned regressions.

By addressing these points, the PR will be more complete, easier to review, and better aligned with the NuttX contribution guidelines.

[yamt]

@xiaoxiang781216
can you please consider to merge this for now, until an alternative like #14827 gets ready?
it's a bad idea to leave things broken for weeks.