txnkv: introduce ballast object to batch-get worker

[zyguan]

Batch-get may spawn multiple goroutines to fetch keys. There is notable cost on stack growing.

After some analysis, it shows the batch-get worker requires 8KB stack size currently, that is, the stack will grow to 4KB, then 8KB for each worker groutine. To reduce the overhead, we can introduce a ballast object to let the stack grow to 8KB directly at the begining of the goroutine.

The benchmark results show that it helps improve performance.

workload	threads	baseline	this PR	diff
batchget	15	19450.06	19548.51	0.51%
batchget	30	24507.40	24914.81	1.66%
batchget	60	26797.00	27451.43	2.44%

BTW, I also tried to use the internal goroutine pool of kvstore, but the test result is not good, the overhead of selectgo is significant, and default 128 pool size is not enough to prevent new goroutines.

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: cfzjywxk, crazycs520

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [cfzjywxk,crazycs520]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[ti-chi-bot]

[LGTM Timeline notifier]

Timeline:

• 2024-12-24 04:01:21.515633967 +0000 UTC m=+1534271.604436515: ☑️ agreed by cfzjywxk.
• 2024-12-24 04:39:29.161587168 +0000 UTC m=+1536559.250389710: ☑️ agreed by crazycs520.
• 2024-12-24 08:30:53.525275697 +0000 UTC m=+1550443.614078233: ✖️🔁 reset by zyguan.
• 2024-12-24 10:00:07.935927534 +0000 UTC m=+1555798.024730073: ☑️ agreed by cfzjywxk.
• 2024-12-25 10:06:53.03591573 +0000 UTC m=+1642603.124718270: ☑️ agreed by crazycs520.