fix: allocate dedicated heap kernel stack for PID 0 (idle task)

PID 0 previously used the boot stack from assembly (_stack_top),
which is not heap-managed. This caused two issues:
- TSS esp0 was not updated when switching to PID 0 (kernel_stack
  was NULL, so the guard in schedule() skipped the update)
- If PID 0 were ever reaped or its stack freed, it would corrupt
  memory since the boot stack is not a kmalloc'd block

Now process_init() allocates a 4KB kernel stack via kmalloc and
sets TSS esp0 to its top, matching the pattern used by all other
processes.

diff --git a/src/kernel/scheduler.c b/src/kernel/scheduler.c
index 766c49db521d333ada79b58c976aefcb34aa622c..e6cc36e43287bad0048d57126b2d63fb51252f31 100644 (file)
--- a/src/kernel/scheduler.c
+++ b/src/kernel/scheduler.c
@@ -632,15 +632,25 @@ void process_init(void) {
     kernel_proc->tls_base = 0;
     kernel_proc->clear_child_tid = NULL;
 
+    /* Allocate a dedicated kernel stack for PID 0 on the heap.
+     * This avoids using the boot stack (which is not heap-managed)
+     * and ensures kfree safety and proper TSS esp0 updates. */
+    void* kstack0 = kmalloc(4096);
+    if (!kstack0) {
+        spin_unlock_irqrestore(&sched_lock, flags);
+        uart_print("[SCHED] OOM allocating PID 0 kernel stack.\n");
+        for (;;) hal_cpu_idle();
+        __builtin_unreachable();
+    }
+    kernel_proc->kernel_stack = (uint32_t*)kstack0;
+
     current_process = kernel_proc;
     ready_queue_head = kernel_proc;
     ready_queue_tail = kernel_proc;
     kernel_proc->next = kernel_proc;
     kernel_proc->prev = kernel_proc;
 
-    // Best effort: set esp0 to current stack until we have a dedicated kernel stack for PID 0
-    uintptr_t cur_esp = hal_cpu_get_stack_pointer();
-    hal_cpu_set_kernel_stack(cur_esp);
+    hal_cpu_set_kernel_stack((uintptr_t)kstack0 + 4096);
 
     spin_unlock_irqrestore(&sched_lock, flags);
 }