What problem does this PR solve?

Issue Number: close #49692

Problem Summary:

• IndexNestedLoopHashJoin.panicErr is not thread-safe

What changed and how does it work?

• make IndexNestedLoopHashJoin.panicErr can not be read until it has been initialized.
• add failpoint TestIssue49692 to make sure the context done branch can be executed.

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No need to test

  • I checked and no code files have been changed.

Manual test:

• Remove

  tidb/pkg/executor/join/index_lookup_hash_join.go

  Lines 204 to 209 in 4a1a771

  	if !e.KeepOuterOrder {
  	e.resultCh <- &indexHashJoinResult{err: err}
  	} else {
  	task := &indexHashJoinTask{err: err}
  	e.taskCh <- task
  	}

  and make sure

  tidb/pkg/executor/join/index_lookup_hash_join.go

  Lines 302 to 306 in 4a1a771

  	failpoint.Inject("TestIssue49692", func() {
  	for !e.panicErr.Load() {
  	runtime.Gosched()
  	}
  	})

  can be triggered.

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

None