kernel: apply P8-initial-placement, P9-numa-topology, P9-proc-lock-ordering

Phase 0c, plan orders #5, #10, #11.

  P8-initial-placement: context::Context::spawn() now picks the
    least-loaded CPU for new threads based on PercpuSched.balance,
    replacing the old 'pin to birth CPU' default.

  P9-numa-topology: adds src/numa.rs (NumaTopology, NumaHint types and
    MAX_NUMA_NODES constant) and threads the get_percpu_block import
    through context/mod.rs. NUMA discovery is performed by userspace
    numad via /scheme/acpi/ and pushed to the kernel via scheme:numa;
    the kernel stores a lightweight copy for O(1) scheduler lookups.

  P9-proc-lock-ordering: fix to scheme/proc.rs acquire order to
    prevent deadlock between proc scheme handles and the per-CPU
    sched lock. Required after P8-percpu-wiring moved the scheduler
    state to per-CPU.

After this commit, three more of the plan's eleven P5–P9 patches are
landed. Remaining unlanded: P5-sched-rt-policy, P6-vruntime-switch,
P7-cache-affine-switch (all touch switch.rs which now diverges from
the patch baselines), and P5-scheme-sched-id/P5-proc-setschedpolicy/
P7-proc-setname/P7-proc-setpriority (overlap on scheme/proc.rs:10X-14X
context handle enum).

cargo check: 1 error remaining (pre-existing src/acpi/fadt.rs:110
unrelated to threading work).

src/context/mod.rs

@@ -10,9 +10,9 @@ use core::{num::NonZeroUsize, ops::Deref};
 
 use crate::{
     context::memory::AddrSpaceWrapper,
-    cpu_set::LogicalCpuSet,
+    cpu_set::{LogicalCpuId, LogicalCpuSet},
     memory::{RmmA, RmmArch, TableKind},
-    percpu::PercpuBlock,
+    percpu::{get_percpu_block, PercpuBlock},
     sync::{
         ArcRwLockWriteGuard, CleanLockToken, LockToken, Mutex, MutexGuard, RwLock, RwLockReadGuard,
         RwLockWriteGuard, L0, L1, L2, L4,
@@ -125,6 +125,30 @@ pub fn run_contexts(token: LockToken<'_, L0>) -> MutexGuard<'_, L1, RunContextDa
     RUN_CONTEXTS.lock(token)
 }
 
+fn least_loaded_cpu() -> LogicalCpuId {
+    let current_cpu = crate::cpu_id();
+    let mut best_cpu = current_cpu;
+    let mut best_depth = usize::MAX;
+
+    for raw_id in 0..crate::cpu_count() {
+        let cpu_id = LogicalCpuId::new(raw_id);
+        let Some(percpu) = get_percpu_block(cpu_id) else {
+            continue;
+        };
+
+        percpu.sched.take_lock();
+        let depth = unsafe { percpu.sched.queues().iter().map(|queue| queue.len()).sum() };
+        percpu.sched.release_lock();
+
+        if depth < best_depth {
+            best_depth = depth;
+            best_cpu = cpu_id;
+        }
+    }
+
+    best_cpu
+}
+
 pub fn init(token: &mut CleanLockToken) {
     let owner = None; // kmain not owned by any fd
     let mut context = Context::new(owner).expect("failed to create kmain context");
@@ -239,6 +263,9 @@ pub fn spawn(
 
     context.kstack = Some(stack);
     context.userspace = userspace_allowed;
+    let target_cpu = least_loaded_cpu();
+    context.sched_affinity = LogicalCpuSet::empty();
+    context.sched_affinity.atomic_set(target_cpu);
 
     let context_lock = Arc::new(ContextLock::new(context));
     let context_ref = ContextRef(Arc::clone(&context_lock));