Implement WFI for Idle Loops and Enforce Shadow Stack Alignment

Optimized system reliability and interrupt latency by implementing wfi in the idle loop
and enforcing 16-byte alignment for the system stack.

Details:
1. Reliability (WFI) Implementation: Modified tx_thread_schedule.S to include the
wfi (Wait For Interrupt) instruction in the idle loop.
The scheduler now enters low-power sleep mode when no threads are ready,
instead of busy-waiting, reducing significant power consumption.
2. Interrupt Latency (Shadow Stack Alignment): Updated tx_initialize_low_level.S to
enforce 16-byte alignment on the system stack pointer (sp) before saving it to
_tx_thread_system_stack_ptr.
Ensures strict adherence to the RISC-V ABI and prevents misaligned access faults
or performance penalties during context switches to the interrupt stack.

Author: Winstonllllai

ports/risc-v32/gnu/src/tx_initialize_low_level.S

@@ -71,6 +71,7 @@ __tx_free_memory_start:
     .global  _tx_initialize_low_level
     .weak    _tx_initialize_low_level
 _tx_initialize_low_level:
+    andi    sp, sp, -16                             // Align stack pointer to 16 bytes
     la      t0, _tx_thread_system_stack_ptr         // Pickup address of system stack ptr
     sw      sp, 0(t0)                               // Save system stack pointer
 

ports/risc-v32/gnu/src/tx_thread_schedule.S

@@ -78,7 +78,11 @@ _tx_thread_schedule:
     la      t0, _tx_thread_execute_ptr                  // Pickup address of execute ptr
 _tx_thread_schedule_loop:
     LOAD    t1, 0(t0)                                   // Pickup next thread to execute
-    beqz    t1, _tx_thread_schedule_loop                // If NULL, wait for thread to execute
+    bnez    t1, _tx_thread_ready                        // If not NULL, break loop
+    wfi                                                 // Wait for interrupt
+    j       _tx_thread_schedule_loop                    // Wait for interrupt
+
+_tx_thread_ready:
 
     /* }
     while(_tx_thread_execute_ptr == TX_NULL);  */