This is an automated cherry-pick of #46055

What problem does this PR solve?

Issue Number: close #46033

Problem Summary:

Normally, TiDB converts timestamp values to UTC before writing them to the storage, and converts them back when reading. For adding index, we

1. Read the table records from storage.
2. Encode the index records.
3. Write/Ingest the index records to storage.

In step 1, the datum returned by extractDatumByOffsets() is in UTC timezone. It is converted to UTC again in step 2. As a result, index timestamp is converted twice.

Unistore DAG request decoding:

TiKV DAG request decoding:
https://github.com/tikv/tikv/blob/14dd8cce96e064a029e58de8cf832c35d44e09a5/components/tidb_query_datatype/src/expr/ctx.rs#L86-L94

For Unistore and TiKV, the time zone name is always considered first, and the time zone offset second.

Thus, the way to buildDAGPB is wrong:

We assigned both name and offset to UTC and -0600 respectively. UTC is kept but -0600 is discarded during parsing the DAG request.

What is changed and how it works?

Assign the time zone offset in buildDAGPB, and leave time zone name to be empty.

// Zone computes the time zone in effect at time t, returning the abbreviated
// name of the zone (such as "CET") and its offset in seconds east of UTC.
func (t Time) Zone() (name string, offset int) {
	name, offset, _, _, _ = t.loc.lookup(t.unixSec())
	return
}

Check List

Tests

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

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