gc checkpoint don't advance when new gc process triggered but current gc is not finished

## Bug Report

Please answer these questions before submitting your issue. Thanks!

### 1. Minimal reproduce step (Required)

asyncFunc.Add(func() {
updateData(ctx1, db, threads, tableNum, insertNum+1, recordSize)
})
asyncFunc.Add(func() {
insertDeleteData(ctx1, db, threads, tableNum, recordSize)
})
asyncFunc.Add(func() {
selectForUpdate(ctx1, db, threads, tableNum, insertNum)
})

//gc default cf
asyncFunc.Add(func() {
updateData(ctx1, db, threads, tableNum, insertNum, recordSize)
})
recordSize = 1024
asyncFunc.Add(func() {
insertDeleteData(ctx1, db, threads, tableNum, recordSize)
})
asyncFunc.Add(func() {
selectForUpdate(ctx1, db, threads, tableNum, insertNum)
})

//delete ranges
asyncFunc.Add(func() {
deleteRanges(ctx1, db, ddlThread, 10, 10000)
})

//check
asyncFunc.Add(func() {
checkSafePoint(ctx1, ctx, db, safePointCheckInterval)
})

### 2. What did you expect to see? (Required)
gc checkpoint can advance
### 3. What did you see instead (Required)
don't advance
### 4. What is your TiDB version? (Required)

1e5b929e16fe5892cdfab18a1ef0268790766060