[BUG] M25/M125 Pause behavior #20969

qwewer0 · 2021-02-01T18:16:33Z

Bug Description

Pausing an SD or serial print has unexpected results.

Test results:

SD printing
- Pause called from the gcode file or gcode sent manually through serial (USB)
  - M25/M125
    1. Print is paused
    2. Printhead is parked
    3. The following three gcode line after M25 is executed on the current (parked) z height
    4. Display shows "Parking...", and is unresponsive, but not frozen (temperature changed correctly)
    5. Connecting to the printer (Pronterface) and sending M24 resumes the pause procedure
    6. Display shows the resume options (Purge more, Continue)
    7. By selecting Continue, the printhead returns to the position where the pause occurred, and continues the print, but skips the already executed three gcodes
  - M25 P0/M125 P0
    Same as M25/M125
  - M25 P1/M125 P1
    1. Print is paused
    2. Printhead is parked
    3. Display shows "Press Button to resume print"
    4. After pressing the button, the display shows the resume options (Purge more, Continue)
    5. By selecting Continue, the printhead returns to the position where the pause occurred, and continues the print
Serial (USB) printing (via Pronterface) or gcode sent manually through serial (USB)
- Pause called from the gcode file
  - M25/M125
    1. Print is paused
    2. Printhead is parked
    3. Display shows "Press Button to resume print"
    4. After pressing the button, the display shows the resume options (Purge more, Continue)
    5. By selecting Continue, the printhead returns to the position where the pause occurred, and continues the print
  - M25 P0/M125 P0
    Same as M25/M125
  - M25 P1/M125 P1
    Same as M25/M125

When the "Press Button to resume print" message is shown on the display sending M24 to the printer has no effect.

Configuration

Printer: Ender 3
Board: SKR Mini E3 v1.2
Firmware: STM32F103RC_btt, Bugfix-2.0.x (c74f972)
Configuration files: Configuration.zip

Requirements

SDSUPPORT
NOZZLE_PARK_FEATURE
ADVANCED_PAUSE_FEATURE
PARK_HEAD_ON_PAUSE
Display

Steps to Reproduce

Start a print through SD card or serial (USB) with M25/M125, M25 P0/M125 P0, M25 P1/M125 P1 in it or send it manually via serial (USB) in mid print
Observe the pausing procedure
Try sending M24 to the printer and/or use the display prompt to resume the printing
Observe the resuming procedure

Expected behavior:

A print paused with M25/M125 or M25 P0/M125 P0 should not show display prompt (unless the code explicitly says so) and wait for M24, which should resume the print (without any display interaction?)
A print paused with M25 P1/M125 P1 should show display prompt (unless the code explicitly says otherwise) and be able to resume the print through the display (and with M24 too?)

Actual behavior:

See "Test results" above, but in short:

An SD print paused with M25/M125 or M25 P0/M125 P0 doesn't show display prompt and waits for M24 (as it should), but after sending M24 to the printer the display is still needed to select "Continue" in order to resume the print
A serial print paused with M25 P0/M125 P0 shows display prompt and the print can be resumed through the display (but not with M24(?))
An SD or serial print paused with M25 P1/M125 P1 shows display prompt and the print can be resumed through the display (as it should) (but not with M24(?))

Additional Information

This issue is connected to #19870

The text was updated successfully, but these errors were encountered:

jmc5113 · 2021-02-06T21:05:28Z

+1 I'm just noticing the same issue :-/

qwewer0 · 2021-02-06T21:10:28Z

+1 I'm just noticing the same issue :-/

Could you share your configuration files?

jmc5113 · 2021-02-06T21:12:18Z

Configuration.zip

qwewer0 · 2021-02-06T21:15:39Z

@jmc5113 Could you try if the issue is still present on the latest bugfix-2.0.x?

jmc5113 · 2021-02-06T22:16:21Z

same thing happened on bugfix-2.0.x.

was that expected to fix it?

jmc5113 · 2021-02-06T22:17:43Z

From the dbg output it looks like extra commands are being executed before it actually pauses...

echo:M117 Place weights in slots and resume printing
echo:M125
echo:busy: processing
echo:busy: processing
echo:busy: processing
echo:busy: processing
echo:busy: processing
echo:busy: processing
echo:busy: processing
echo:busy: processing
X:10.00 Y:280.00 Z:5.00 E:-2.00 Count X:1600 Y:44800 Z:2000
echo:G1 X155.529 Y138.560 F12000.000
echo:G11
echo:G92 E0
X:155.53 Y:138.56 Z:5.00 E:0.00 Count X:1600 Y:44800 Z:2000

jmc5113 · 2021-02-06T22:18:26Z

and when it does pause, I can't use the display to unpause it. I have to use M24 from serial console.

jmc5113 · 2021-02-06T22:20:34Z

Also when I put in M24, it initiates the purge sequence, which I do not need as I'm only trying to pause.

qwewer0 · 2021-02-06T22:21:17Z

same thing happened on bugfix-2.0.x.

was that expected to fix it?

It's best to test the latest bugfix-2.0.x, not to work for nothing.

jmc5113 · 2021-02-06T22:30:39Z

just discovered...

M125 P1 // will move to park and pause and show the correct prompt on the screen.

https://marlinfw.org/docs/gcode/M125.html

qwewer0 · 2021-03-09T06:07:08Z

Still here.

thinkyhead · 2021-03-09T06:27:30Z

NOTE: M25 is not supported as a command run from SD because it isn't a command that blocks the queue. It simply tells the machine to stop processing commands from the open SD print-job.

qwewer0 · 2021-03-09T06:33:14Z

NOTE: M25 is not supported as a command run from SD because it isn't a command that blocks the queue. It simply tells the machine to stop processing commands from the open SD print-job.

M25 and M125 same with PARK_HEAD_ON_PAUSE enabled, so then is it only M25 that isn't suppose to be used or M125 too?

Marlin/Marlin/src/gcode/sd/M24_M25.cpp

Lines 91 to 97 in 2f1fd4b

    
           void GcodeSuite::M25() { 
        
             #if ENABLED(PARK_HEAD_ON_PAUSE) 
        
               M125(); 
        
             #else

@thinkyhead Thank you for looking into this.

thinkyhead · 2021-03-09T06:34:17Z

M125 is the supported method. M25 as you point out is not guaranteed.

thinkyhead · 2021-03-09T06:39:03Z

Was there some point in the past when these commands were behaving better? If so, how recently did this bad behavior start to appear?

The reason I mention that M25 is non-blocking is that it only runs when it goes to the front of the queue, so there are always going to be some commands in the queue behind it. Therefore it will not behave as one expects when used in a file printing from SD.

qwewer0 · 2021-03-09T06:44:04Z

Was there some point in the past when these commands were behaving better? If so, how recently did this bad behavior start to appear?

If I remember it correctly, it behaved like this from the start when I switched from 8bit 1.1.x to 32bit 2.0.x at around 2019. Dec.

thinkyhead · 2021-03-09T06:44:31Z

When the "Press Button to resume print" message is shown on the display sending M24 to the printer has no effect.

The prompt must be dismissed by a host prompt or by M108 in this case. The M24 command is historically still a queued command and Marlin makes no special allowance to execute it out of sequence.

qwewer0 · 2021-03-09T06:46:00Z

The prompt must be dismissed by a host prompt or by M108 in this case. The M24 command is historically still a queued command and Marlin makes no special allowance to execute it out of sequence.

Is there a better way to test the pause behavior to get more information?

thinkyhead · 2021-03-09T06:58:17Z

I don't quite understand the question.

For testing, we can disregard how M25 behaves when used from within a G-code file printing from SD.
We always want M25 to return right away and to never block the queue.
Since M25 can now insert a park, it might be prudent to call queue.exaust() in the park function.
The M125 command is supported as a command within a G-code stream and can block the queue. So it's ok to call it from an SD print. When called otherwise … well …

/**
 * M125: Store current position and move to parking position.
 *       Called on pause (by M25) to prevent material leaking onto the
 *       object. On resume (M24) the head will be moved back and the
 *       print will resume.
 *
 *       When not actively SD printing, M125 simply moves to the park
 *       position and waits, resuming with a button click or M108.
 *       Without PARK_HEAD_ON_PAUSE the M125 command does nothing.
 *
 *    L<linear> = Override retract Length
 *    X<pos>    = Override park position X
 *    Y<pos>    = Override park position Y
 *    Z<linear> = Override Z raise
 *
 *  With an LCD menu:
 *    P<bool>   = Always show a prompt and await a response
 */

thinkyhead · 2021-03-09T07:02:37Z

Since M25 can now insert a park, it might be prudent to call queue.exaust() in the park function.

That said, queue.exaust() should be considered dangerous because it could in rare and unusual cases itself invoke more pause commands or dwells…

thinkyhead · 2021-03-09T07:33:17Z

I see that if code was added to GCodeQueue::get_sdcard_commands (or even GCodeQueue::RingBuffer::commit_command) to check for the M25 command and call card.pauseSDPrint() then it would prevent any more commands from the SD card being added to the queue until the print was resumed.

qwewer0 mentioned this issue Mar 4, 2021

[BUG] M25/M125 without P1 parameter should wait for M24 #19870

Closed

github-actions bot added the stale-closing-soon label Mar 9, 2021

MarlinFirmware deleted a comment from github-actions bot Mar 9, 2021

thinkyhead added Bug: Potential ? Needs: Work More work is needed and removed stale-closing-soon labels Mar 9, 2021

thinkyhead mentioned this issue Mar 11, 2021

Pause SD queue early on M25 #21317

Merged

qwewer0 mentioned this issue May 8, 2021

fixes #21813 #21828

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[BUG] M25/M125 Pause behavior #20969

[BUG] M25/M125 Pause behavior #20969

qwewer0 commented Feb 1, 2021

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021 •

edited

jmc5113 commented Feb 6, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021 •

edited

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

thinkyhead commented Mar 9, 2021 •

edited

[BUG] M25/M125 Pause behavior #20969

[BUG] M25/M125 Pause behavior #20969

Comments

qwewer0 commented Feb 1, 2021

Bug Description

Configuration

Requirements

Steps to Reproduce

Additional Information

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

jmc5113 commented Feb 6, 2021

qwewer0 commented Feb 6, 2021 • edited

jmc5113 commented Feb 6, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021 • edited

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

qwewer0 commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

thinkyhead commented Mar 9, 2021

thinkyhead commented Mar 9, 2021 • edited

qwewer0 commented Feb 6, 2021 •

edited

thinkyhead commented Mar 9, 2021 •

edited

thinkyhead commented Mar 9, 2021 •

edited