Implement Mutexes and supporting syscalls, cleanup/optimize scheduler

kernel/amd64/mm.c

@@ -240,15 +240,11 @@ uintptr_t mm_alloc_user_pd_phys (void) {
 
 /* Reload after map/unmap operation was performed. This function does the TLB shootdown. */
 void mm_reload (void) {
-  spin_lock (&mm_lock);
-
   struct limine_mp_response* mp = limine_mp_request.response;
 
   for (size_t i = 0; i < mp->cpu_count; i++) {
     amd64_lapic_ipi (mp->cpus[i]->lapic_id, TLB_SHOOTDOWN);
   }
-
-  spin_unlock (&mm_lock);
 }
 
 bool mm_validate (struct pd* pd, uintptr_t vaddr, uint32_t flags) {

kernel/amd64/proc.c

@@ -50,7 +50,7 @@ struct proc* proc_from_elf (uint8_t* elf_contents) {
   struct proc_resource* ustk_r =
       proc_create_resource (proc, ustk_rid, PR_MEM, RV_PRIVATE, (void*)&ustk_mem_init);
   if (ustk_r == NULL) {
-    kstk_r->ops.cleanup (kstk_r);
+    kstk_r->ops.cleanup (proc, kstk_r);
     free (kstk_r);
     free (proc);
     return NULL;

kernel/amd64/smp.c

@@ -8,7 +8,7 @@
 #include <libk/string.h>
 #include <limine/requests.h>
 #include <mm/liballoc.h>
-#include <sync/rw_spin_lock.h>
+#include <sync/spin_lock.h>
 #include <sys/debug.h>
 #include <sys/smp.h>
 #include <sys/syscall.h>
@@ -27,7 +27,7 @@ struct cpu* cpu_make (void) {
   struct cpu* cpu = &cpus[id];
 
   memset (cpu, 0, sizeof (*cpu));
-  cpu->lock = RW_SPIN_LOCK_INIT;
+  cpu->lock = SPIN_LOCK_INIT;
   cpu->id = id;
   cpu->self = cpu;
 

kernel/amd64/smp.h

@@ -7,7 +7,7 @@
 #include <libk/rbtree.h>
 #include <libk/std.h>
 #include <proc/proc.h>
-#include <sync/rw_spin_lock.h>
+#include <sync/spin_lock.h>
 
 #define CPUS_MAX 32
 
@@ -30,7 +30,7 @@ struct cpu {
     atomic_int nesting;
   } irq_ctx;
 
-  rw_spin_lock_t lock;
+  spin_lock_t lock;
 
   struct rb_node_link* proc_run_q;
   struct proc* proc_current;