KAFKA-17862: [buffer pool] corruption during buffer reuse from the pool

[gongxuanzhang]

issue see https://issues.apache.org/jira/browse/KAFKA-17862

🔍 Problem Summary

When an expired batch is still part of an in-flight request, we
prematurely release the ByteBuffer back to the BufferPool. This leads to
two critical issues:

1. Expiration does not prevent the in-flight request from being sent.
2. The expired batch’s ByteBuffer is deallocate to the pool too early.
   It may be re-allocated for another producer batch while still being
   referenced by the in-flight request, potentially causing data
   corruption.

We can tolerate Issue 1, but Issue 2 is critical — we cannot allow it to
happen.

Therefore, we remove the expiration handling of ProducerBatch before
send, and instead defer the ByteBuffer deallocation to the response
handling logic.

[github-actions]

A label of 'needs-attention' was automatically added to this PR in order to raise the
attention of the committers. Once this issue has been triaged, the triage label
should be removed to prevent this automation from happening again.

[github-actions]

A label of 'needs-attention' was automatically added to this PR in order to raise the
attention of the committers. Once this issue has been triaged, the triage label
should be removed to prevent this automation from happening again.

[github-actions]

A label of 'needs-attention' was automatically added to this PR in order to raise the
attention of the committers. Once this issue has been triaged, the triage label
should be removed to prevent this automation from happening again.