…f func_name

Resolve every defined symbol's libbpf-compatible file offset once per
library (st_value - sh_addr + sh_offset) and attach uprobes by offset,
replacing the per-probe func_name lookup that reparsed the ELF each time.

Symbol resolution parses each allocator ELF independently, so fan it out
across cores with rayon before the serial attach step. Cuts offset
resolution for all discovered allocators from ~190ms to ~29ms locally.

Closing a uprobe link fd blocks on two RCU grace periods in the kernel,
serialized per link (~1,200 links = ~50s teardown). Concurrent waiters
share grace periods, so closing from 32 threads scales near-linearly.

The tracker must be detached explicitly in track_command: the IPC thread
holds an Arc clone, so the tracker is never dropped before process::exit
and the kernel would close every link fd serially during exit. Tracking
is disabled right after the child exits so draining and artifact writes
happen before the expensive detach.