Use thread local storage for thread index

This patch deprecates stack-based thread identification,
Also removes requirement that thread stacks are adjacent.

Finally, possibly annoying for some folks, it renames
all occurences of cpu_index and cpu_number with thread
index. Using word "cpu" is misleading here as thread can
be migrated ti different CPU, and also it is not related
to linux cpu index.

Change-Id: I68cdaf661e701d2336fc953dcb9978d10a70f7c1
Signed-off-by: Damjan Marion <damarion@cisco.com>

2 files changed, 16 insertions, 16 deletions

diff --git a/src/vnet/dpo/lookup_dpo.c b/src/vnet/dpo/lookup_dpo.c
index e94e871cd92..97ad0a44cce 100644
--- a/src/vnet/dpo/lookup_dpo.c
+++ b/src/vnet/dpo/lookup_dpo.c
@@ -266,7 +266,7 @@ lookup_dpo_ip4_inline (vlib_main_t * vm,
                        int table_from_interface)
 {
     u32 n_left_from, next_index, * from, * to_next;
-    u32 cpu_index = os_get_cpu_number();
+    u32 thread_index = vlib_get_thread_index();
     vlib_combined_counter_main_t * cm = &load_balance_main.lbm_to_counters;
 
     from = vlib_frame_vector_args (from_frame);
@@ -407,10 +407,10 @@ lookup_dpo_ip4_inline (vlib_main_t * vm,
 	    vnet_buffer(b1)->ip.adj_index[VLIB_TX] = dpo1->dpoi_index;
 
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi0, 1,
+		(cm, thread_index, lbi0, 1,
 		 vlib_buffer_length_in_chain (vm, b0));
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi1, 1,
+		(cm, thread_index, lbi1, 1,
 		 vlib_buffer_length_in_chain (vm, b1));
 
 	    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
@@ -511,7 +511,7 @@ lookup_dpo_ip4_inline (vlib_main_t * vm,
 	    vnet_buffer(b0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
 
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi0, 1,
+		(cm, thread_index, lbi0, 1,
 		 vlib_buffer_length_in_chain (vm, b0));
 
 	    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
@@ -606,7 +606,7 @@ lookup_dpo_ip6_inline (vlib_main_t * vm,
 {
     vlib_combined_counter_main_t * cm = &load_balance_main.lbm_to_counters;
     u32 n_left_from, next_index, * from, * to_next;
-    u32 cpu_index = os_get_cpu_number();
+    u32 thread_index = vlib_get_thread_index();
 
     from = vlib_frame_vector_args (from_frame);
     n_left_from = from_frame->n_vectors;
@@ -749,10 +749,10 @@ lookup_dpo_ip6_inline (vlib_main_t * vm,
 	    vnet_buffer(b1)->ip.adj_index[VLIB_TX] = dpo1->dpoi_index;
 
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi0, 1,
+		(cm, thread_index, lbi0, 1,
 		 vlib_buffer_length_in_chain (vm, b0));
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi1, 1,
+		(cm, thread_index, lbi1, 1,
 		 vlib_buffer_length_in_chain (vm, b1));
 
 	    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
@@ -853,7 +853,7 @@ lookup_dpo_ip6_inline (vlib_main_t * vm,
 	    vnet_buffer(b0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
 
 	    vlib_increment_combined_counter
-		(cm, cpu_index, lbi0, 1,
+		(cm, thread_index, lbi0, 1,
 		 vlib_buffer_length_in_chain (vm, b0));
 
 	    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED))
@@ -930,7 +930,7 @@ lookup_dpo_mpls_inline (vlib_main_t * vm,
                        int table_from_interface)
 {
     u32 n_left_from, next_index, * from, * to_next;
-    u32 cpu_index = os_get_cpu_number();
+    u32 thread_index = vlib_get_thread_index();
     vlib_combined_counter_main_t * cm = &load_balance_main.lbm_to_counters;
 
     from = vlib_frame_vector_args (from_frame);
@@ -994,7 +994,7 @@ lookup_dpo_mpls_inline (vlib_main_t * vm,
             vnet_buffer(b0)->ip.adj_index[VLIB_TX] = dpo0->dpoi_index;
 
             vlib_increment_combined_counter
-                (cm, cpu_index, lbi0, 1,
+                (cm, thread_index, lbi0, 1,
                  vlib_buffer_length_in_chain (vm, b0));
 
 	    if (PREDICT_FALSE(b0->flags & VLIB_BUFFER_IS_TRACED)) 
diff --git a/src/vnet/dpo/replicate_dpo.c b/src/vnet/dpo/replicate_dpo.c
index a9f334be7cf..e25ceae91fe 100644
--- a/src/vnet/dpo/replicate_dpo.c
+++ b/src/vnet/dpo/replicate_dpo.c
@@ -627,7 +627,7 @@ replicate_inline (vlib_main_t * vm,
     vlib_combined_counter_main_t * cm = &replicate_main.repm_counters;
     replicate_main_t * rm = &replicate_main;
     u32 n_left_from, * from, * to_next, next_index;
-    u32 cpu_index = os_get_cpu_number();
+    u32 thread_index = vlib_get_thread_index();
 
     from = vlib_frame_vector_args (frame);
     n_left_from = frame->n_vectors;
@@ -657,12 +657,12 @@ replicate_inline (vlib_main_t * vm,
             rep0 = replicate_get(repi0);
 
             vlib_increment_combined_counter(
-                cm, cpu_index, repi0, 1,
+                cm, thread_index, repi0, 1,
                 vlib_buffer_length_in_chain(vm, b0));
 
-	    vec_validate (rm->clones[cpu_index], rep0->rep_n_buckets - 1);
+	    vec_validate (rm->clones[thread_index], rep0->rep_n_buckets - 1);
 
-	    num_cloned = vlib_buffer_clone (vm, bi0, rm->clones[cpu_index], rep0->rep_n_buckets, 128);
+	    num_cloned = vlib_buffer_clone (vm, bi0, rm->clones[thread_index], rep0->rep_n_buckets, 128);
 
 	    if (num_cloned != rep0->rep_n_buckets)
 	      {
@@ -673,7 +673,7 @@ replicate_inline (vlib_main_t * vm,
 
             for (bucket = 0; bucket < num_cloned; bucket++)
             {
-                ci0 = rm->clones[cpu_index][bucket];
+                ci0 = rm->clones[thread_index][bucket];
                 c0 = vlib_get_buffer(vm, ci0);
 
                 to_next[0] = ci0;
@@ -700,7 +700,7 @@ replicate_inline (vlib_main_t * vm,
 		    vlib_get_next_frame (vm, node, next_index, to_next, n_left_to_next);
 		  }
             }
-	    vec_reset_length (rm->clones[cpu_index]);
+	    vec_reset_length (rm->clones[thread_index]);
         }
 
         vlib_put_next_frame (vm, node, next_index, n_left_to_next);