RepRapFirmware 2.0 with RTOS in development

dc42

More progress yesterday and today. It turned out that there are a lot of chunks of code that are not thread-safe, especially in the older code - which is understandable, because it was written before using a RTOS was envisaged. Anyway, I've now made all the output channels thread safe, also the tool list and the mechanism used to extract print data from GCode files.

I've started including a binary of the latest RTOS build for the Duet WiFi and Ethernet in the github commits. It's at https://github.com/dc42/RepRapFirmware/blob/v2-dev/EdgeRelease/Duet2CombinedFirmware.bin. If you like taking risks, follow that link and click Download. I've included a couple of bug fixes in this one too:

Fixed issue with G1 S1 Ennn on delta printers
The FileInfo parser now has a timeout if called from DWC, to prevent the connection being lost when getting info from a large file with a small SD card cluster size. You will no longer get disconnected, but you probably won't get object height or filament required information.

It's compatible with version 1.21 of iap4e.bin, DuetWiFiServer.bin and DuetWebControl.zip.

cgcafer

Hi David,
This morning i updated, after that i can't connect via wifi and fan2 (now it is connected to laser head) it was flashed one in 3-5 second.

I was busy so after that i reinstall to regular version (1.12). If you want to diagnostic i can try it again.
Thanks
Cafer

dc42

I presume you mean 1.21 not 1.12.

What was the problem with Fan2?

The RTOS build is working on my delta, except that the first time I booted it up this morning the wifi was off and I had to send M552 S1 from PanelDue to enable it. I have a new build almost ready, but I'll do some tests powering up from cold to see if that problem occurs again.

cgcafer

Yes you are right 1.21
laser led turned on and off(like flashing) in 3-5 seconds
I was tried M552 S1 but no success.

cobretti

Just updated my D-Bot and all good so far. No issues after update with warm and cold restarts and a successful small print.
Edit: Running a Duet Wifi

DaveA

Just ran a test print on my Anycubic delta. Printed fine. I did have a problem when I realized I had selected the wrong file to print and attempted to pause/cancel. The pause seemed to hang everything. The bed was still heating but I couldn't kill the print. I needed to power cycle the printer and re-connect to start again. Other than that everything seems good.
This is with a Duet Ethernet.

dc42

@DaveA, thanks for that. I'll add pause/resume to the list of things to be tested before the next release.

Danal

Downloaded and installed the 2.0(RTOS)alpha1 (2018-04-02b3). Uploaded and installed via the DWC dialogs. Everything worked, and it came back online quite quickly.

This is a Duet WiFi (reads as Duet WiFi 1.02 or later) with nothing attached except power. Home switch configurations are set so that a G28 succeeds (NC, I think). Everything that I can test with nothing attached looks good.

I'll put this duet in a printer over the next few days so I can test more.

Danal

Vin seems to read a fraction high, as compared to an Eventek benchtop power supply, and a Mastech DVM. They agree to within 0.02.

At 12.0 V on the power supply, 12.02 on the DVM, DWC shows 12.2.
At 24.0 V on the power supply, 24.02 on the DVM, DWC shows 24.4.

This may be entirely within normal tolerance, I did not check it on the "old" firmware.

dc42

The tolerance is +/-4.5% but it should typically be within +/-2%.

dc42

I've just pushed a new version to github. This one runs the network processes as a separate task, and doing that has restored the file upload speed to about what is was before. Further improvement should be possible in future.

Pause seems to be working. I suspect it was a stack overflow, as I has to increase one of the stack sizes to fix another issue.

If you try this version, please keep an eye on the stack sizes using M122 (but see below), and let me know if any of the "stack rem" figures fall below 100.

Known bug in this version: if you send M122 from DWC, it sometimes disconnects because of a JSON error, and you can't reconnect DWC without rebooting the Duet. M122 from USB is OK. I'll fix this in the next release.

Danal

2.0(RTOS)alpha1 (2018-04-03b2) loaded successfully through DWC.

Successfully simulated a print, and uploaded a file at the same time. About 454 K/Sec.

Danal

After the sim and upload in parallel.

[[language]]
8:51:45 AMM122
=== Diagnostics ===
=== RTOS ===
Static ram used: 26196
Dynamic ram used: 95116
Recycled dynamic ram: 1568
Handler stack ram used: 280
Never used ram: 7912
Task NETWORK: state 0 stack rem 676
Task HEAT: state 2 stack rem 172
Task MAIN: state 1 stack rem 428
=== Platform ===
RepRapFirmware for Duet 2 WiFi/Ethernet version 2.0(RTOS)alpha1 running on Duet WiFi 1.02 or later
Board ID: 08DGM-9568A-F23SD-6JTDJ-3SD6S-TTNMF
Last reset 00:10:41 ago, cause: software
Last software reset at 2018-04-02 21:44, reason: Hard fault, spinning module GCodes, available RAM 3488 bytes (slot 0)
Software reset code 0x4033 HFSR 0x40000000, CFSR 0x00008200, ICSR 0x0441f803, BFAR 0x41010000, SP 0x20001fb4
Stack: 0040a643 00432e5c 01070000 ffffffff 20002524 ffffffff 0040ac2d 004384e4 a1070000 200091f0 00443477 20002008 ffffffff 00000000 00000000 200091f0 00000000 ffffffff a0000001 2000219c 0044352b 20002020 20002064 20007cb0
Used output buffers: 10 of 32 (23 max)
Error status: 0
Free file entries: 10
SD card 0 detected, interface speed: 20.0MBytes/sec
SD card longest block write time: 168.9ms
MCU temperature: min 37.2, current 38.1, max 39.1
Supply voltage: min 12.1, current 12.2, max 12.3, under voltage events: 0, over voltage events: 0
Driver 0: standstill, SG min/max not available
Driver 1: standstill, SG min/max not available
Driver 2: standstill, SG min/max not available
Driver 3: standstill, SG min/max not available
Driver 4: standstill, SG min/max not available
Date/time: 2018-04-03 08:51:45
Slowest main loop (seconds): 0.219054; fastest: 0.000065
=== Move ===
MaxReps: 0, StepErrors: 0, LaErrors: 0, FreeDm: 240, MinFreeDm 240, MaxWait: 0ms, Underruns: 0, 0
Scheduled moves: 0, completed moves: 0
Bed compensation in use: none
Bed probe heights: 0.000 0.000 0.000 0.000 0.000
=== Heat ===
Bed heaters = 0 -1 -1 -1, chamberHeaters = -1 -1
=== GCodes ===
Segments left: 0
Stack records: 1 allocated, 0 in use
Movement lock held by null
http is idle in state(s) 0
telnet is idle in state(s) 0
file is idle in state(s) 0
serial is idle in state(s) 0
aux is idle in state(s) 0
daemon is idle in state(s) 0
queue is idle in state(s) 0
autopause is idle in state(s) 0
Code queue is empty.
=== Network ===
Responder states: HTTP(1) HTTP(0) HTTP(0) HTTP(0) FTP(0) Telnet(0) Telnet(0)
HTTP sessions: 1 of 8
- WiFi -
Network state is running
WiFi module is connected to access point 
Failed messages: pending 0, notready 0, noresp 0
WiFi firmware version 1.21
WiFi MAC address 2c:3a:e8:0a:ec:ba
WiFi Vcc 3.38, reset reason Turned on by main processor
WiFi flash size 4194304, free heap 15952
WiFi IP address 192.168.86.110
WiFi signal strength -45dBm, reconnections 0, sleep mode modem
Socket states: 2 0 0 0 0 0 0 0
=== Expansion ===

dc42

Danal, thanks.

Which time zone are you in? The software reset data says:

Last reset 00:10:41 ago, cause: software
Last software reset at 2018-04-02 21:44, reason: Hard fault, spinning module GCodes, available RAM 3488 bytes (slot 0)

but I don't think they can refer to the same reset unless your computer clock is more than 16 hours behind mine.

EDIT: now that I think about it, the automatic reset after uploading new firmware won't update the software reset data, so that explains it.

cgcafer

I have another duet wifi and i updated RTOS firmware via DWC. After that duet wifi continuously reset

NOTE : Duet wifi board connected usb port and no any other connection (any motor, any sensor etc only duet wifi board v1.02)

DaveA

Things are not going so well after the latest build.
I've been getting a number of disconnects due to syntax errors in the JSON. I wasn't getting these in the previous build.
I've not been able to print Benchy with the new build. It will start printing and then some random time later it will just stop with a 100% complete message. I've not been able to print beyond 40-50 layers before stopping.
I reinstalled the 1.21 build and Benchy is now almost complete without error. I didn't keep the previous RTOS build so I can't try it.

I did try to cancel a print and it seemed to work but it had to complete the bed warming before it would it cancel. Maybe this is expected behavior.

S1lencer

Hi, i think i found a kink in the new RTOS Firmware, if you try to run an autocalibration "M302 H1 S240" DWC said the usual stuff but not long after that the Board reboots.

dc42

I've just pushed a new version to github. I've fixed some issues with the C library not being reentrant, which were causing occasional crashes.

@S1lencer, thanks for your report. Did you mean M303? I haven't tried heater tuning with v2 yet, and as the heater control is now a separate process it may be that the stack for that process is too small to handle heater tuning. I'll add this to the bug list.

@DaveA, I've seen that DWC will disconnect after a while with a JSON parse error too. Do you have a PanelDue? I suspect that this may not happen without one. Putting the PanelDue on the Setup page (so that it stops polling) may avoid it.

I've fixed a couple of issues unrelated to RTOS:

If multiple devices are connected to a Duet via the network and DWC, and one of them goes to sleep without disconnecting first, this will be recognised after 8 seconds so the other device will no longer receive repeated messages.
M114 reports user coordinates first and machine coordinates at the end.

S1lencer

Jep did mean M303

DaveA

I do have a PanelDue. I disconnected it and had no further disconnects. Didn't try putting it in Setup.
Tried three prints today. All Benchys.
First print completed successfully.
Second print never starting printing before claiming 100% done and returning to idle. Don't know if heating completed.
Third print completed successfully.
I didn't reset between these prints. Just picked the print again button.