Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Some Test to investigate input stream errors #14

Closed
wants to merge 12 commits into from
Closed

Some Test to investigate input stream errors #14

wants to merge 12 commits into from

Conversation

AnHardt
Copy link
Owner

@AnHardt AnHardt commented Jan 29, 2016

And two experimental solutions.

If you are interested, read commit by commit.

AnHardt added 11 commits January 27, 2016 12:52
@AnHardt
Test if RX-buffer is full
700c3d6
@AnHardt
Provoke some Errors
0afa1e5 
The easiest way to provoke some errors is to acticate:
#define AUTO_BED_LEVELING_FEATURE
#define Z_PROBE_END_SCRIPT "G1 Z10 F12000\nG1 X15 Y330\nG1 Z0.5\nG1 Z10"
in Configuration.h

and send some g-code-file including a G29.

With RepetierHostSettings_1.png,
i runned killercone.gcode an got Test_1.txt, sent with Repetier host.
With RepetierHostSettings_2.png,
i runned killercone.gcode an got Test_2.txt, sent with Repetier host.

Search for "error:".

We get some of our newly introduced
Error:Serial: incoming chars rejected - ring buffer full
followd by
Error:checksum mismatch, Last Line: xx
and or
Error:Line Number is not Last Line Number+1, Last Line: xx

Retesting with our new error deactivated gives the same errors without the new one.

With the second RH-settings we get the errors a bit later - with the second G29.
Thats caused by a different buffer filling strategy.

With Pronterface i get verry similar errors as with RepetierHostSettings_2.png

Lets take a clother view on Marlins buffer usage.
@AnHardt
Activate ADVANCED_OK
319d031 
ADVANCED_OK provides some output for the buffer usage.

Now we get:
ok N34 P10 B1
With N = Linenumber, P = free entries in the planner buffer, B = free entrys in the buffer for preparsed g-codes.

Sadly no information about the RX-buffer.

A test with RepetierHostSettings_2.png gives Test_3.txt.

First search for "error:" - the errors seem to occure much later.
Then search for "N40" - Why do we OK that so often? Likely these are the commands we encued with G29. We have not had linenumbers for them. But G29 was sent with N26? Someting is wrong.
@AnHardt
Extend ADVANCED_OK
e473500 
by R = free places in the RX-buffer
and the output of the g.code line we just finished.

See Test_4.txt
The original ADVANCED_OK sends a N-number not matching the actual line.
The number send corosponds with the line just put __in__ the comand buffer, not the one we finished.
@AnHardt
Repair N in ADVANCED_OK
f30c8fc 
Send the correct line number.
If no N was in the command - non't send an N.

See Test_5.txt
Now the linenumbers do make sense.

Let's have a look on the R values. Until the first G29 they are 128 - unused, everything fitted into the B-buffer.
Behind the first G29 we begin to use the R-buffer.
Somewhen behind the second G29 the R-buffer is completely full and we get errors.
@AnHardt
Experimental: Don't send "ok" when no line number was received
9d417da 
else send "echo:".

Test with RepetierHostSettings_2.png -> Test_6.txt
Test with RepetierHostSettings_1.png -> Test_7.txt
Test with Pronterface Watch ON -> Test_8.txt
Test with Pronterface Watch OFF -> Test_9.txt

No errors - in all 4 cases!

Let's have a closer lock.

a)
Test_6.txt line 33-36: N skips from N2 to N4 -> that's an M105 - RepetieHost does not log M105 but sends a line number with it.
Test_7.txt line 32-33: the same
Test_8.txt line 163-165; Pronterface sends M105 without N

that does not matter because M105 sends it's own ok - not with ADVANCED_OK

b)
Looking at the end of the test-files

The P-buffer is well used in all tests.
The B-buffer is only used with RH (RepetierHostSettings_1.png Test_7.txt) mode and when lines are injected by enqueuecommands_P (Test_6.txt 82-108, Test_8.txt 61-105, Test_9.txt 55-91)
The R-buffer has about the same use as the B-buffer.

Pronterface and RH in mode RepetierHostSettings_2.png use a strategy to send a new line for every "ok" they receive.
RH in mode RepetierHostSettings_1.png tries to count how many chars are in the buffer. So if a long line was "ok"ed it may send two short lines (f.e. Test_7.txt 740-760).

All in all the use of the buffers is bad in all cases, but at least we are save now - not loosing input.

Can we do something to improve the buffer usage?
@AnHardt
Early ok
a062c61
@AnHardt
Complete RX-buffer full test
28e69ac
@AnHardt
In case of a protocol error print the malicious string
3d73624
@AnHardt
Add kill() for unattended tests.
fd9e3cb
@AnHardt
Some more tests
9eab49c 
Tries to speed up.

Preliminary Result:

OK for 115200bd!!!
Likely a race condition between checkRx() and store_char() at 250000bd.
Possibly a race condition between those two and get_command().

Replacing MarlinSerial with HardwareSerial does not help for 250000bd.
For 115200bd HardwareSerial is not better but biger.
@AnHardt AnHardt mentioned this pull request Jan 29, 2016
Closed
AnHardt pushed a commit that referenced this pull request Feb 24, 2016
Merge pull request #14 from MarlinFirmware/RCBugFix
34cde1b 
Rc bug fix
@AnHardt AnHardt closed this Mar 23, 2016
@AnHardt AnHardt deleted the debug_input branch March 23, 2016 15:34
AnHardt pushed a commit that referenced this pull request Aug 25, 2019
@ETE-Design @thinkyhead
# This is a combination of 18 commits.
de122c9 
# This is the 1st commit message:

Add Support for ADIMLab Granty

# The commit message #2 will be skipped:

# Update Configuration.h

# The commit message #3 will be skipped:

# Add files via upload
#
# Add Support for ADIMLab Granty

# The commit message #4 will be skipped:

# Create Configuration.h

# The commit message #5 will be skipped:

# Delete Configuration.h

# The commit message #6 will be skipped:

# Create Test.h

# The commit message #7 will be skipped:

# Add files via upload

# The commit message #8 will be skipped:

# Delete Test.h

# The commit message #9 will be skipped:

# Delete Configuration.h

# The commit message #10 will be skipped:

# Delete Configuration_adv.h

# The commit message #11 will be skipped:

# Add files via upload
#
# Add ADIMLab Pins

# The commit message #12 will be skipped:

# Add files via upload
#
# Add Bootscreen

# The commit message #13 will be skipped:

# Update Configuration.h
#
# Add Bootscreen

# The commit message #14 will be skipped:

# Update pins.h
#
# Add ADIMLab Board

# The commit message #15 will be skipped:

# Update boards.h
#
# Add ADIMLab Granty

# The commit message #16 will be skipped:

# Update Makefile
#
# Add ADIMLab Granty

# The commit message #17 will be skipped:

# Update boards.h

# The commit message #18 will be skipped:

# Update Makefile
AnHardt pushed a commit that referenced this pull request Aug 25, 2019
@thinkyhead
ADIMLab Granty v1 / v2 (MarlinFirmware#15017)
20fc66f 
* # This is a combination of 18 commits.
# This is the 1st commit message:

Add Support for ADIMLab Granty

# The commit message #2 will be skipped:

# Update Configuration.h

# The commit message #3 will be skipped:

# Add files via upload
#
# Add Support for ADIMLab Granty

# The commit message #4 will be skipped:

# Create Configuration.h

# The commit message #5 will be skipped:

# Delete Configuration.h

# The commit message #6 will be skipped:

# Create Test.h

# The commit message #7 will be skipped:

# Add files via upload

# The commit message #8 will be skipped:

# Delete Test.h

# The commit message #9 will be skipped:

# Delete Configuration.h

# The commit message #10 will be skipped:

# Delete Configuration_adv.h

# The commit message #11 will be skipped:

# Add files via upload
#
# Add ADIMLab Pins

# The commit message #12 will be skipped:

# Add files via upload
#
# Add Bootscreen

# The commit message #13 will be skipped:

# Update Configuration.h
#
# Add Bootscreen

# The commit message #14 will be skipped:

# Update pins.h
#
# Add ADIMLab Board

# The commit message #15 will be skipped:

# Update boards.h
#
# Add ADIMLab Granty

# The commit message #16 will be skipped:

# Update Makefile
#
# Add ADIMLab Granty

# The commit message #17 will be skipped:

# Update boards.h

# The commit message #18 will be skipped:

# Update Makefile

* Add Support for ADIMLab Granty

* Optimize bootscreen

* Update Makefile

Updated Description

* Update boards.h

Updated Description

* Update pins_HJC2560C_REV2.h

Updated Description

* Update Configuration.h

Added Author + Custom_Machine_Name

* Create _Bootscreen.h

Add Bootscreen

* Add files via upload

Add ADIMLab Granty v1

* Update boards.h

Add ADIMLab Granty v1

* Update Makefile

Add ADIMLab Granty v1

* Update pins_HJC2560C_REV2.h

* Consolidate Granty configs

* config updates
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers
No reviews
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
1 participant
@AnHardt