feat: overhaul userspace heap and virtual memory

kernel/src/Sched/Scheduler.hpp

@@ -33,6 +33,9 @@ namespace Sched {
     // failed bound check is recoverable (SYS_ALLOC returns null); allowing the
     // heap to cross this boundary would silently replace library PTEs.
     static constexpr uint64_t UserHeapLimit = 0x60000000ULL;
+    static constexpr uint64_t UserHeapPages =
+        (UserHeapLimit - UserHeapBase) / 0x1000ULL;
+    static constexpr uint64_t UserHeapBitmapWords = (UserHeapPages + 63) / 64;
     // Surface mappings use reusable, fixed-size per-process slots instead of
     // consuming heap VA forever.  64 * 32 MiB occupies [0x7000000000,
     // 0x7080000000), comfortably below the main user stack.
@@ -191,6 +194,7 @@ namespace Sched {
     // reservation is serialized across sibling threads and bounded below the
     // fixed shared-library mappings.
     bool ReserveUserHeapRange(int primarySlot, uint64_t size, uint64_t& outVa);
+    void ReleaseUserHeapRange(int primarySlot, uint64_t va, uint64_t size);
 
     // Terminate the currently executing thread. If this is the main thread,
     // the entire process exits (equivalent to ExitProcess).
@@ -258,5 +262,8 @@ namespace Sched {
 
     // Per-process allocated page count (tracked by Heap syscalls, separate from Process struct)
     inline uint64_t g_allocatedPages[MaxProcesses] = {};
+    // One bit per page in the bounded userspace heap. Unlike heapNext, this
+    // makes virtual ranges reusable after unmap and failed reservations.
+    inline uint64_t g_userHeapPageMap[MaxProcesses][UserHeapBitmapWords] = {};
 
 }