From d5eb8cd9169411877805f896450b986bf1730cc5 Mon Sep 17 00:00:00 2001
From: Markus Hitter <mah@jump-ing.de>
Date: Sun, 27 Nov 2016 20:54:28 +0100
Subject: [PATCH] DDA: avoid looking up the movebuffer array.

As we have mb_tail_dda now, that's no longer necessary. Using
something like movebuffer[mb_tail] is more expensive than
dereferencing mb_tail_dda directly.

This is the first time we see a stepping performance improvement
since introducing mb_tail_dda. 13 clock cycles faster on the
slowest step, which is 9 cycles faster than before that
introduction.

Binary size also a nice 94 bytes down.

  ATmega sizes               '168   '328(P)   '644(P)     '1280
  Program:  19270 bytes      135%       63%       31%       15%
     Data:   2179 bytes      213%      107%       54%       27%
   EEPROM:     32 bytes        4%        2%        2%        1%

  short-moves.gcode statistics:
  LED on occurences: 888.
  LED on time minimum: 218 clock cycles.
  LED on time maximum: 395 clock cycles.
  LED on time average: 249.051 clock cycles.

  smooth-curves.gcode statistics:
  LED on occurences: 23648.
  LED on time minimum: 237 clock cycles.
  LED on time maximum: 438 clock cycles.
  LED on time average: 272.216 clock cycles.

  triangle-odd.gcode statistics:
  LED on occurences: 1636.
  LED on time minimum: 237 clock cycles.
  LED on time maximum: 395 clock cycles.
  LED on time average: 262.572 clock cycles.
---
 clock.c         | 10 +++++-----
 dda.c           |  4 ++--
 dda_queue.c     |  6 +++---
 gcode_process.c | 14 +++++++-------
 4 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/clock.c b/clock.c
index b01b58f..01a8ab1 100644
--- a/clock.c
+++ b/clock.c
@@ -116,11 +116,11 @@ static void clock_250ms(void) {
 
 			// target position
       sersendf_P(PSTR("Dst: %lq,%lq,%lq,%lq,%lu\n"),
-                 movebuffer[mb_tail].endpoint.axis[X],
-                 movebuffer[mb_tail].endpoint.axis[Y],
-                 movebuffer[mb_tail].endpoint.axis[Z],
-                 movebuffer[mb_tail].endpoint.axis[E],
-                 movebuffer[mb_tail].endpoint.F);
+                 mb_tail_dda->endpoint.axis[X],
+                 mb_tail_dda->endpoint.axis[Y],
+                 mb_tail_dda->endpoint.axis[Z],
+                 mb_tail_dda->endpoint.axis[E],
+                 mb_tail_dda->endpoint.F);
 
 			// Queue
 			print_queue();
diff --git a/dda.c b/dda.c
index 92cf576..9ca1fe6 100644
--- a/dda.c
+++ b/dda.c
@@ -940,7 +940,7 @@ void dda_clock() {
 
 /// update global current_position struct
 void update_current_position() {
-	DDA *dda = &movebuffer[mb_tail];
+  DDA *dda = mb_tail_dda;
   enum axis_e i;
 
   static const axes_uint32_t PROGMEM steps_per_m_P = {
@@ -950,7 +950,7 @@ void update_current_position() {
     STEPS_PER_M_E
   };
 
-  if (dda->live) {
+  if (dda != NULL) {
     uint32_t axis_steps, axis_um;
 
     for (i = X; i < AXIS_COUNT; i++) {
diff --git a/dda_queue.c b/dda_queue.c
index 93ac376..7f987c5 100644
--- a/dda_queue.c
+++ b/dda_queue.c
@@ -58,8 +58,8 @@ uint8_t queue_full() {
 /// Take a step or go to the next move.
 void queue_step() {
 
-  if (movebuffer[mb_tail].live) {
-    dda_step(&movebuffer[mb_tail]);
+  if (mb_tail_dda != NULL) {
+    dda_step(mb_tail_dda);
 	}
 
   /**
@@ -67,7 +67,7 @@ void queue_step() {
 
     This needs no atomic protection, because we're in an interrupt already.
   */
-  if ( ! movebuffer[mb_tail].live) {
+  if ( ! mb_tail_dda->live) {
     if (mb_tail != mb_head) {
       mb_tail = MB_NEXT(mb_tail);
       mb_tail_dda = &(movebuffer[mb_tail]);
diff --git a/gcode_process.c b/gcode_process.c
index 5c185ca..73922c1 100644
--- a/gcode_process.c
+++ b/gcode_process.c
@@ -611,15 +611,15 @@ void process_gcode_command() {
 
         if (DEBUG_POSITION && (debug_flags & DEBUG_POSITION)) {
           sersendf_P(PSTR("Endpoint: X:%ld,Y:%ld,Z:%ld,E:%ld,F:%lu,c:%lu}\n"),
-                     movebuffer[mb_tail].endpoint.axis[X],
-                     movebuffer[mb_tail].endpoint.axis[Y],
-                     movebuffer[mb_tail].endpoint.axis[Z],
-                     movebuffer[mb_tail].endpoint.axis[E],
-                     movebuffer[mb_tail].endpoint.F,
+                     mb_tail_dda->endpoint.axis[X],
+                     mb_tail_dda->endpoint.axis[Y],
+                     mb_tail_dda->endpoint.axis[Z],
+                     mb_tail_dda->endpoint.axis[E],
+                     mb_tail_dda->endpoint.F,
                      #ifdef ACCELERATION_REPRAP
-                       movebuffer[mb_tail].end_c
+                       mb_tail_dda->end_c
                      #else
-                       movebuffer[mb_tail].c
+                       mb_tail_dda->c
                      #endif
           );
           print_queue();