BR: Restore pre-split fails with "split key exceeds limit" when restoring to TiDB X (BYOC)

## Bug Report

### 1. Minimal reproduce step (Required)

1. Backup a large-scale TiDB cluster with high region density.
2. Restore to **TiDB X (BYOC)** using BR.
3. During **Restore Pre-split**, BR generates > 1024 split keys for a single range/table.
4. BR sends all keys in one RPC batch, which the **rfstore** module rejects (hard limit: 1024).

### 2. What did you expect to see? (Required)

BR should automatically **chunk split key requests** into multiple batches (e.g., max 1024 per RPC) to ensure compatibility with cloud storage engine constraints without manual tuning.

### 3. What did you see instead (Required)

Restore fails with "split key count exceeds limit."

Currently, users must use "hacky" workarounds to bypass this:
* `--merge-region-size-bytes=1159296000`
* `--merge-region-key-count=115000000`

These parameters force oversized regions and cause post-restore performance imbalances.

### 4. What is your TiDB version? (Required)

v8.5.0