ManuelW/grbl-LPC-CoreXY
1
0
Fork 0
Code Issues Pull Requests Releases 1 Wiki Activity

Refactoring and lots of bug fixes. Updated homing cycle.

Browse Source 
WARNING: There are still some bugs to be worked out. Please use caution
if you test this firmware.

- Feed holds work much better, but there are still some failure
conditions that need to be worked out. This is the being worked on
currently and a fix is planned to be pushed next.

- Homing cycle refactoring: Slight adjustment of the homing cycle to
allow for limit pins to be shared by different axes, as long as the
shared limit pins are not homed on the same cycle. Also, removed the
LOCATE_CYCLE portion of the homing cycle configuration. It was
redundant.

- Limit pin sharing: (See above). To clear up one or two limit pins for
other IO, limit pins can now be shared. For example, the Z-limit can be
shared with either X or Y limit pins, because it’s on a separate homing
cycle. Hard limit will still work exactly as before.

- Spindle pin output fixed. The pins weren’t getting initialized
correctly.

- Fixed a cycle issue where streaming was working almost like a single
block mode. This was caused by a problem with the spindle_run() and
coolant_run() commands and issuing an unintended planner buffer sync.

- Refactored the cycle_start, feed_hold, and other runtime routines
into the runtime command module, where they should be handled here
only. These were redundant.

- Moved some function calls around into more appropriate source code
modules.

- Fixed the reporting of spindle state.
This commit is contained in:
Sonny Jeon
2014-02-09 10:46:34 -07:00
parent cc9afdc195
commit 50fbc6e297
20 changed files with 579 additions and 529 deletions
Download Patch File Download Diff File Expand all files Collapse all files

251
protocol.c
View File

@ -1,5 +1,5 @@
/*
protocol.c - controls Grbl execution procedures
protocol.c - controls Grbl execution protocol and procedures
Part of Grbl
Copyright (c) 2011-2014 Sungeun K. Jeon
@ -24,6 +24,7 @@
#include "settings.h"
#include "protocol.h"
#include "gcode.h"
#include "planner.h"
#include "stepper.h"
#include "motion_control.h"
#include "report.h"
@ -32,95 +33,6 @@
static char line[LINE_BUFFER_SIZE]; // Line to be executed. Zero-terminated.
// Executes run-time commands, when required. This is called from various check points in the main
// program, primarily where there may be a while loop waiting for a buffer to clear space or any
// point where the execution time from the last check point may be more than a fraction of a second.
// This is a way to execute runtime commands asynchronously (aka multitasking) with grbl's g-code
// parsing and planning functions. This function also serves as an interface for the interrupts to
// set the system runtime flags, where only the main program handles them, removing the need to
// define more computationally-expensive volatile variables. This also provides a controlled way to
// execute certain tasks without having two or more instances of the same task, such as the planner
// recalculating the buffer upon a feedhold or override.
// NOTE: The sys.execute variable flags are set by any process, step or serial interrupts, pinouts,
// limit switches, or the main program.
void protocol_execute_runtime()
{
// Reload step segment buffer
st_prep_buffer();
if (sys.execute) { // Enter only if any bit flag is true
uint8_t rt_exec = sys.execute; // Avoid calling volatile multiple times
// System alarm. Everything has shutdown by something that has gone severely wrong. Report
// the source of the error to the user. If critical, Grbl disables by entering an infinite
// loop until system reset/abort.
if (rt_exec & (EXEC_ALARM | EXEC_CRIT_EVENT)) {
sys.state = STATE_ALARM; // Set system alarm state
// Critical event. Only hard/soft limit errors currently qualify.
if (rt_exec & EXEC_CRIT_EVENT) {
report_alarm_message(ALARM_LIMIT_ERROR);
report_feedback_message(MESSAGE_CRITICAL_EVENT);
bit_false(sys.execute,EXEC_RESET); // Disable any existing reset
do {
// Nothing. Block EVERYTHING until user issues reset or power cycles. Hard limits
// typically occur while unattended or not paying attention. Gives the user time
// to do what is needed before resetting, like killing the incoming stream. The
// same could be said about soft limits. While the position is not lost, the incoming
// stream could be still engaged and cause a serious crash if it continues afterwards.
} while (bit_isfalse(sys.execute,EXEC_RESET));
// Standard alarm event. Only abort during motion qualifies.
} else {
// Runtime abort command issued during a cycle, feed hold, or homing cycle. Message the
// user that position may have been lost and set alarm state to enable the alarm lockout
// to indicate the possible severity of the problem.
report_alarm_message(ALARM_ABORT_CYCLE);
}
bit_false(sys.execute,(EXEC_ALARM | EXEC_CRIT_EVENT));
}
// Execute system abort.
if (rt_exec & EXEC_RESET) {
sys.abort = true; // Only place this is set true.
return; // Nothing else to do but exit.
}
// Execute and serial print status
if (rt_exec & EXEC_STATUS_REPORT) {
report_realtime_status();
bit_false(sys.execute,EXEC_STATUS_REPORT);
}
// Initiate stepper feed hold
if (rt_exec & EXEC_FEED_HOLD) {
// !!! During a cycle, the segment buffer has just been reloaded and full. So the math involved
// with the feed hold should be fine for most, if not all, operational scenarios.
st_feed_hold(); // Initiate feed hold.
bit_false(sys.execute,EXEC_FEED_HOLD);
}
// Reinitializes the stepper module running state and, if a feed hold, re-plans the buffer.
// NOTE: EXEC_CYCLE_STOP is set by the stepper subsystem when a cycle or feed hold completes.
if (rt_exec & EXEC_CYCLE_STOP) {
st_cycle_reinitialize();
bit_false(sys.execute,EXEC_CYCLE_STOP);
}
if (rt_exec & EXEC_CYCLE_START) {
st_cycle_start(); // Issue cycle start command to stepper subsystem
if (bit_istrue(settings.flags,BITFLAG_AUTO_START)) {
sys.auto_start = true; // Re-enable auto start after feed hold.
}
bit_false(sys.execute,EXEC_CYCLE_START);
}
}
// Overrides flag byte (sys.override) and execution should be installed here, since they
// are runtime and require a direct and controlled interface to the main stepper program.
}
// Directs and executes one line of formatted input from protocol_process. While mostly
// incoming streaming g-code blocks, this also directs and executes Grbl internal commands,
// such as settings, initiating the homing cycle, and toggling switch states.
@ -141,20 +53,27 @@ static void protocol_execute_line(char *line)
status = system_execute_line(line);
} else {
// Everything else is gcode. Send to g-code parser!
// Everything else is gcode. Send to g-code parser! Block if in alarm mode.
if (sys.state == STATE_ALARM) { status = STATUS_ALARM_LOCK; }
else { status = gc_execute_line(line); }
// TODO: Separate the parsing from the g-code execution. Need to re-write the parser
// completely to do this. First parse the line completely, checking for modal group
// errors and storing all of the g-code words. Then, send the stored g-code words to
// a separate g-code executor. This will be more in-line with actual g-code protocol.
status = gc_execute_line(line);
// TODO: Clean up the multi-tasking workflow with the execution of commands. It's a
// bit complicated and patch-worked. Could be made simplier to understand.
}
report_status_message(status);
}
void protocol_process()
/*
GRBL MAIN LOOP:
*/
void protocol_main_loop()
{
// ------------------------------------------------------------
// Complete initialization procedures upon a power-up or reset.
@ -218,15 +137,149 @@ void protocol_process()
}
}
protocol_execute_runtime(); // Runtime command check point.
if (sys.abort) { return; } // Bail to main() program loop to reset system.
// If there are no more characters in the serial read buffer to be processed and executed,
// this indicates that g-code streaming has either filled the planner buffer or has
// completed. In either case, auto-cycle start, if enabled, any queued moves.
mc_auto_cycle_start();
protocol_auto_cycle_start();
protocol_execute_runtime(); // Runtime command check point.
if (sys.abort) { return; } // Bail to main() program loop to reset system.
}
return; /* Never reached */
}
// Executes run-time commands, when required. This is called from various check points in the main
// program, primarily where there may be a while loop waiting for a buffer to clear space or any
// point where the execution time from the last check point may be more than a fraction of a second.
// This is a way to execute runtime commands asynchronously (aka multitasking) with grbl's g-code
// parsing and planning functions. This function also serves as an interface for the interrupts to
// set the system runtime flags, where only the main program handles them, removing the need to
// define more computationally-expensive volatile variables. This also provides a controlled way to
// execute certain tasks without having two or more instances of the same task, such as the planner
// recalculating the buffer upon a feedhold or override.
// NOTE: The sys.execute variable flags are set by any process, step or serial interrupts, pinouts,
// limit switches, or the main program.
void protocol_execute_runtime()
{
uint8_t rt_exec = sys.execute; // Copy to avoid calling volatile multiple times
if (rt_exec) { // Enter only if any bit flag is true
// System alarm. Everything has shutdown by something that has gone severely wrong. Report
// the source of the error to the user. If critical, Grbl disables by entering an infinite
// loop until system reset/abort.
if (rt_exec & (EXEC_ALARM | EXEC_CRIT_EVENT)) {
sys.state = STATE_ALARM; // Set system alarm state
// Critical event. Only hard/soft limit errors currently qualify.
if (rt_exec & EXEC_CRIT_EVENT) {
report_alarm_message(ALARM_LIMIT_ERROR);
report_feedback_message(MESSAGE_CRITICAL_EVENT);
bit_false(sys.execute,EXEC_RESET); // Disable any existing reset
do {
// Nothing. Block EVERYTHING until user issues reset or power cycles. Hard limits
// typically occur while unattended or not paying attention. Gives the user time
// to do what is needed before resetting, like killing the incoming stream. The
// same could be said about soft limits. While the position is not lost, the incoming
// stream could be still engaged and cause a serious crash if it continues afterwards.
} while (bit_isfalse(sys.execute,EXEC_RESET));
// Standard alarm event. Only abort during motion qualifies.
} else {
// Runtime abort command issued during a cycle, feed hold, or homing cycle. Message the
// user that position may have been lost and set alarm state to enable the alarm lockout
// to indicate the possible severity of the problem.
report_alarm_message(ALARM_ABORT_CYCLE);
}
bit_false(sys.execute,(EXEC_ALARM | EXEC_CRIT_EVENT));
}
// Execute system abort.
if (rt_exec & EXEC_RESET) {
sys.abort = true; // Only place this is set true.
return; // Nothing else to do but exit.
}
// Execute and serial print status
if (rt_exec & EXEC_STATUS_REPORT) {
report_realtime_status();
bit_false(sys.execute,EXEC_STATUS_REPORT);
}
// Execute a feed hold with deceleration, only during cycle.
if (rt_exec & EXEC_FEED_HOLD) {
// !!! During a cycle, the segment buffer has just been reloaded and full. So the math involved
// with the feed hold should be fine for most, if not all, operational scenarios.
if (sys.state == STATE_CYCLE) {
sys.state = STATE_HOLD;
st_update_plan_block_parameters();
st_prep_buffer();
sys.auto_start = false; // Disable planner auto start upon feed hold.
}
bit_false(sys.execute,EXEC_FEED_HOLD);
}
// Execute a cycle start by starting the stepper interrupt begin executing the blocks in queue.
if (rt_exec & EXEC_CYCLE_START) {
if (sys.state == STATE_QUEUED) {
sys.state = STATE_CYCLE;
st_prep_buffer(); // Initialize step segment buffer before beginning cycle.
st_wake_up();
if (bit_istrue(settings.flags,BITFLAG_AUTO_START)) {
sys.auto_start = true; // Re-enable auto start after feed hold.
} else {
sys.auto_start = false; // Reset auto start per settings.
}
}
bit_false(sys.execute,EXEC_CYCLE_START);
}
// Reinitializes the cycle plan and stepper system after a feed hold for a resume. Called by
// runtime command execution in the main program, ensuring that the planner re-plans safely.
// NOTE: Bresenham algorithm variables are still maintained through both the planner and stepper
// cycle reinitializations. The stepper path should continue exactly as if nothing has happened.
// NOTE: EXEC_CYCLE_STOP is set by the stepper subsystem when a cycle or feed hold completes.
if (rt_exec & EXEC_CYCLE_STOP) {
if (sys.state != STATE_QUEUED) {
sys.state = STATE_IDLE;
}
bit_false(sys.execute,EXEC_CYCLE_STOP);
}
}
// Overrides flag byte (sys.override) and execution should be installed here, since they
// are runtime and require a direct and controlled interface to the main stepper program.
// Reload step segment buffer
if (sys.state & (STATE_CYCLE | STATE_HOLD | STATE_HOMING)) { st_prep_buffer(); }
}
// Block until all buffered steps are executed or in a cycle state. Works with feed hold
// during a synchronize call, if it should happen. Also, waits for clean cycle end.
void protocol_buffer_synchronize()
{
// Check and set auto start to resume cycle after synchronize and caller completes.
if (sys.state == STATE_CYCLE) { sys.auto_start = true; }
while (plan_get_current_block() || (sys.state == STATE_CYCLE)) {
protocol_execute_runtime(); // Check and execute run-time commands
if (sys.abort) { return; } // Check for system abort
}
}
// Auto-cycle start has two purposes: 1. Resumes a plan_synchronize() call from a function that
// requires the planner buffer to empty (spindle enable, dwell, etc.) 2. As a user setting that
// automatically begins the cycle when a user enters a valid motion command manually. This is
// intended as a beginners feature to help new users to understand g-code. It can be disabled
// as a beginner tool, but (1.) still operates. If disabled, the operation of cycle start is
// manually issuing a cycle start command whenever the user is ready and there is a valid motion
// command in the planner queue.
// NOTE: This function is called from the main loop and mc_line() only and executes when one of
// two conditions exist respectively: There are no more blocks sent (i.e. streaming is finished,
// single commands), or the planner buffer is full and ready to go.
void protocol_auto_cycle_start() { if (sys.auto_start) { sys.execute |= EXEC_CYCLE_START; } }