Kernel: Store SchedulerThreadNode directly in Thread

Having the pointer indirection was weird and forced allocations when
adding/binding a new thread :P

kernel/include/kernel/Processor.h

@@ -22,7 +22,6 @@ namespace Kernel
 	class GDT;
 	class IDT;
 	class Scheduler;
-	class SchedulerThreadNode;
 	class Thread;
 
 #if ARCH(x86_64) || ARCH(i686)
@@ -54,8 +53,8 @@ namespace Kernel
 			union
 			{
 				TLBEntry flush_tlb;
-				SchedulerThreadNode* new_thread;
-				SchedulerThreadNode* unblock_thread;
+				Thread* new_thread;
+				Thread* unblock_thread;
 				bool dummy;
 			};
 		};

kernel/include/kernel/Scheduler.h

@@ -41,9 +41,8 @@ namespace Kernel
 		void on_timer_interrupt();
 		void on_yield(YieldRegisters*);
 
-		static BAN::ErrorOr<void> bind_thread_to_processor(Thread*, ProcessorID);
-		// if thread is already bound, this will never fail
-		BAN::ErrorOr<void> add_thread(Thread*);
+		static void bind_thread_to_processor(Thread*, ProcessorID);
+		void add_thread(Thread*);
 
 		void block_current_thread(ThreadBlocker* thread_blocker, uint64_t wake_time_ns, BaseMutex* mutex);
 		void unblock_thread(Thread*);
@@ -68,9 +67,6 @@ namespace Kernel
 
 		class ProcessorID find_least_loaded_processor() const;
 
-		void add_thread(SchedulerThreadNode*);
-		void unblock_thread(SchedulerThreadNode*);
-
 	private:
 		SchedulerQueue m_run_list;
 		SchedulerHeap m_block_list;

kernel/include/kernel/Thread.h

@@ -7,6 +7,7 @@
 #include <kernel/InterruptStack.h>
 #include <kernel/Lock/Mutex.h>
 #include <kernel/Memory/VirtualRange.h>
+#include <kernel/SchedulerThreadNode.h>
 
 #include <LibELF/AuxiliaryVector.h>
 
@@ -192,7 +193,7 @@ namespace Kernel
 		vaddr_t                    m_fsbase               { 0 };
 		vaddr_t                    m_gsbase               { 0 };
 
-		SchedulerThreadNode*       m_scheduler_node       { nullptr };
+		SchedulerThreadNode        m_scheduler_node;
 
 		YieldRegisters             m_yield_registers      { };