txn: fix retried transactions' statements is not inclueded in TIDB_TRX

[longfangsong]

What problem does this PR solve?

Problem Summary:

Currently the retried transactions' statements won't be inclueded in TIDB_TRX.

The reason is:

If we meet an retryable error here:

tidb/session/session.go

Line 701 in 091ae3b

err = s.doCommit(ctx)

note the origin datasource of TIDB_TRX for this transaction is resetted here inside doCommit:

tidb/session/session.go

Line 489 in 091ae3b

s.txn.changeToInvalid()

(and will be resetted again here if retry more than once):

tidb/session/session.go

Line 972 in 091ae3b

s.txn.changeToInvalid()

But during retrying, we use st.Exec(which may restart the transaction with session.Txn) instead of session.ExecStament, and st.Exec will not record statements.

tidb/session/session.go

Line 929 in 091ae3b

_, err = st.Exec(ctx)

What is changed and how it works?

What's Changed:
Collect them during retry.

Check List

Tests

• Unit test
• Integration test (automated-tests#851)
• Manual test
• No code

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory

Documentation

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

Release note

None

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• MyonKeminta
• cfzjywxk

To complete the pull request process, please ask the reviewers in the list to review by filling /cc @reviewer in the comment.
After your PR has acquired the required number of LGTMs, you can assign this pull request to the committer in the list by filling /assign @committer in the comment to help you merge this pull request.

The full list of commands accepted by this bot can be found here.

Reviewer can indicate their review by submitting an approval review.
Reviewer can cancel approval by submitting a request changes review.

[longfangsong]

/cc @MyonKeminta @cfzjywxk
And please also /cc the test case in automated-tests#851

[cfzjywxk]

/merge

[ti-chi-bot]

This pull request has been accepted and is ready to merge.

Commit hash: cceb7a1

[ti-srebot]

cherry pick to release-5.2 in PR #28670

[MyonKeminta]

@longfangsong Does 5.1 branch need this?

[longfangsong]

@longfangsong Does 5.1 branch need this?

Yes, we should cherry-pick this to 5.1 if possible.

However, it seems there are no 5.1 bug triage plan here

[longfangsong]

/label needs-cherry-pick-5.1

[ti-srebot]

cherry pick to release-5.1 in PR #28716

[MyonKeminta]

I just notice that it looks that the onStmtEnd function might be invoked multiple times on a same txn. If that's true, though it's not a problem currently according to the implementation of onStmtEnd, I think maybe we need someway to guarantee onStmtEnd to be reentrant. At least we need to add some comments to notice people to avoid breaking it.