fix(spurd): remove dead code and wire cgroup cleanup

[shiv-tyagi]

Remove superseded functions from executor.rs (job_execution_loop, report_completion, signal_job, run_epilog) and the unused create_server helper from agent_server.rs. Drop the never-read job_id field from RunningJob variants and the test-only which() utility from container.rs.

Wire cleanup_cgroup into the agent server's monitor loop so cgroup directories are properly removed when jobs finish. Clean up all unused imports and prefix intentionally-unused parameters with underscore.

[Copilot]

Pull request overview

This PR removes superseded/dead code in the spurd agent implementation and wires job cgroup cleanup into the agent’s monitoring loop so cgroup directories are removed when jobs finish.

Changes:

• Removed unused helpers/fields (executor loop stubs, RunningJob.job_id, agent_server::create_server, container::which() + tests) and cleaned up unused imports/params.
• Exposed and invoked cgroup cleanup from the agent server’s job monitor loop on job completion.
• Minor housekeeping in GPU/Landlock modules (unused imports, intentionally-unused parameter naming).

Reviewed changes

Copilot reviewed 6 out of 6 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
crates/spurd/src/reporter.rs	Removes unused imports after resource reporting refactors.
crates/spurd/src/landlock.rs	Removes unused warn import.
crates/spurd/src/gpu.rs	Removes unused warn import and marks an unused parameter.
crates/spurd/src/executor.rs	Deletes dead executor-loop code, removes RunningJob.job_id, adds take_cgroup, and makes cleanup_cgroup callable from the agent server.
crates/spurd/src/container.rs	Removes the test-only which() helper and its tests.
crates/spurd/src/agent_server.rs	Drops unused server helper and wires cgroup cleanup into the completion/monitor loop.

Comments suppressed due to low confidence (1)

crates/spurd/src/executor.rs:542

• cleanup_cgroup only reads/kills PIDs from cgroup.procs in the current cgroup. In cgroup v2, cgroup.procs does not include processes in descendant cgroups, so jobs that create nested cgroups can leave processes running and prevent successful removal. Consider using cgroup.kill when available, or recursively walking child cgroups to kill their cgroup.procs before attempting removal.

    // Kill any remaining processes
    if let Ok(pids) = std::fs::read_to_string(cgroup_path.join("cgroup.procs")) {
        for pid_str in pids.lines() {
            if let Ok(pid) = pid_str.trim().parse::<i32>() {
                let _ = signal::kill(Pid::from_raw(pid), Signal::SIGKILL);

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[shiv-tyagi]

Addressed the copilot review comments. Should be okay to merge now. The other things raised by copilot should be done in a separate PR.