conn: add write timeout for conn which receive a kill signal

[AilinKid]

What problem does this PR solve?

Issue Number: close #44009

Problem Summary:

What is changed and how it works?

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.

add write timeout for conn which receive a kill signal

[ti-chi-bot]

[REVIEW NOTIFICATION]

This pull request has been approved by:

• Defined2014
• wshwsh12

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.

[AilinKid]

/merge

[ti-chi-bot]

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

Commit hash: 776747c

[Defined2014]

/hold

[AilinKid]

for loop {
    // drain the select result
    execStmt.Next(ctx, chunk) {
        Next(executor)
        kill singal check
    }
    dump the chunk to conn {
        start := time.now()
        dump chunk to data
        write(data) to conn
        write-resp += time.since(start)
    }
}

##### an explanation for slow-log queries with a long write-resp time:
we couldn't say a statement skip out of all the kill singal check, since Next calling is in 
the unit of a chunk(1024 row). If a statement has a long write-resp in slow log, that means a
accumulated long write-time of these data to conn.


##### an explanation for problem queries that couldn't be killed:
When it comes to queries that can't be killed, the only guess is that we don't have a second 
Next loop check for a kill signal and exits as an interrupt, and the dump time keeps 
getting stuck this time around. 

We need a wakeable write data implementation when a blocking write data has occurred.

[AilinKid]

We have interspersed the Next interface call in the interval of write data with chunk as the unit to realize the detection of kill signal, but the blocking problem of write-data itself is still an important solution to this problem, and default timeout may not be a good solution way of blocking. Considering that this problem only occurs when the network is extremely poor, temporarily close if there is no elegant solution. (chinese explanation: 我们已经在以 chunk 为单元的 write data 的间隔中穿插了 Next 接口调用来实现 kill signal 的检测，但是 write-data 自身的阻塞问题是仍然是这个问题的重要解，default timeout 可能并不是一个好的解除阻塞的方式，考虑到这问题只有在网络极差的情况下才发生，没有优雅解法的情况下暂时close)