Describe the problem

In SQL, traceIDs/spanIDs are currently represented using DInt, which is an int64-backed type. The tracing package uses uint64s. While this isn't a technical issue, it might be confusing to see int64(x) via SQL and uint64(x) in logs.

Expected behavior
SQL and tracing should agree. Either we use BIGINT in SQL or uint64 in tracing. If we do the latter, we have to be a little careful about the migration (we have to stick to uint64 on the wire, in (*Tracer).InjectX/ExtractX).

Additional data / screenshots
Discussed on https://reviewable.io/reviews/cockroachdb/cockroach/59992

Jira issue: CRDB-3163