diff --git a/include/kernel/thread.h b/include/kernel/thread.h
index 99a6c41..0bd286a 100644
--- a/include/kernel/thread.h
+++ b/include/kernel/thread.h
@@ -40,6 +40,11 @@
 
 typedef int (*thread_start_routine)(void *arg);
 
+/* thread local storage */
+enum thread_tls_list {
+	MAX_TLS_ENTRY
+};
+
 #define THREAD_MAGIC 'thrd'
 
 typedef struct thread {
@@ -71,6 +76,9 @@
 	/* return code */
 	int retcode;
 
+	/* thread local storage */
+	uint32_t tls[MAX_TLS_ENTRY];
+
 	char name[32];
 } thread_t;
 
@@ -140,6 +148,19 @@
 static inline void inc_critical_section(void) { critical_section_count++; }
 static inline void dec_critical_section(void) { critical_section_count--; }
 
+/* thread local storage */
+static inline __ALWAYS_INLINE uint32_t tls_get(uint entry)
+{
+	return current_thread->tls[entry];
+}
+
+static inline __ALWAYS_INLINE uint32_t tls_set(uint entry, uint32_t val)
+{
+	uint32_t oldval = current_thread->tls[entry];
+	current_thread->tls[entry] = val;
+	return oldval;
+}
+
 /* wait queue stuff */
 #define WAIT_QUEUE_MAGIC 'wait'
 
diff --git a/kernel/thread.c b/kernel/thread.c
index fc9d9ca..1a23478 100644
--- a/kernel/thread.c
+++ b/kernel/thread.c
@@ -122,6 +122,11 @@
 
 	t->stack_size = stack_size;
 
+	/* inheirit thread local storage from the parent */
+	int i;
+	for (i=0; i < MAX_TLS_ENTRY; i++)
+		t->tls[i] = current_thread->tls[i];
+
 	/* set up the initial stack frame */
 	arch_thread_initialize(t);
 
@@ -456,6 +461,12 @@
 	dprintf(INFO, "\tstack %p, stack_size %zd\n", t->stack, t->stack_size);
 	dprintf(INFO, "\tentry %p, arg %p\n", t->entry, t->arg);
 	dprintf(INFO, "\twait queue %p, wait queue ret %d\n", t->blocking_wait_queue, t->wait_queue_block_ret);
+	dprintf(INFO, "\ttls:");
+	int i;
+	for (i=0; i < MAX_TLS_ENTRY; i++) {
+		dprintf(INFO, " 0x%x", t->tls[i]);
+	}
+	dprintf(INFO, "\n");
 }
 
 void dump_all_threads(void)