V1.0.0 (R853) crash induced by tunnel + logic analyzer condition trigger

I tried to clean up a circuit a bit, but it crashes on start (sometimes on second / third start).

Seen at: WIN10, R853

Situation:

• circuit with mega328: SDA / SCL on MCU connected to tunnel, second instances ot these connected to pull up, third instances connected to logic analyzer
  
• MCU used without firmware
  
• logic analyzer triggers on condition CH1
  

sim1 file is attached.

The crashes do not appear, when no tunnels are used.

Ok... the bug also appears when no tunnel is used.
Maybe it is also a different one. In this case, I had to use some small code:

Code:


int main(void)
{
 DDRD = 0b00000010;
 PORTD = 0b00000010;
 PORTD = 0b00000000;
 PORTD = 0b00000010;
 
 while (1)  {};
}


Then, the analzyer had to trigger on CH1 (per condition). Crash appears on WIN10, R853 - had to start the simulation 3-4 times in order to get the crash. Hex and sim1 attached.

Edit: the "code-based crash" seems somehow to be related to the PortD1, It seems not appear on PortB1.

I can't reproduce the issue, not even in Windows.
Same thing with second example...

EDIT: Let me try downloading it....

No way.. I can't get it to crash.
Even downloading simulide from here.

Are you sure you are using the last files I uploaded?
I reuploaded it the first time because there was a posible bug that looks exactly like this one...

The bug is really strange: It seems to appear more often, when clicking on start fast. I downloaded the simu version and my files again and can reproduce the bug.

See here: https://wiki.mexle.org/_media/2022-01-15_21-01-58.mkv

I had a restart of WIN and WIN updates (and getting a screen recorder) in between. It seems to take now more starts to chrash simulide ..


Edit:
Here a clip of the "code induced" one: https://wiki.mexle.org/_media/2022-01-15_21-15-26.mkv
(this one crashes for me on the first start after flashing)

Yes.. I could reproduce it by clicking really really fast in Windows10 (virtual machine).
No way to reproduce it in Linux (where I can debug it).

I will try to find a possible cause...

Have you also tried the "code induced" one?
This seem to be more reproducable also with low click rate (and currently annoying in my productive code   ).


Strangely, the bug seem to be also depending on the simulation speed. For 0.005% and slower it seems not to be present (simu step 1us). For 0.008% and more it crashes quite regularily on my WIN machine.

Have you also tried the "code induced" one?

Yes, but is not easy to make it crash, I have to click like crazy and only sometimes happens.

Strangely, the bug seem to be also depending on the simulation speed. For 0.005% and slower it seems not to be present (simu step 1us). For 0.008% and more it crashes quite regularily on my WIN machine.

This is a good clue.
I will add some code to not accept clicks until last action is finished.
Hopefully that will solve the issue.

If you have a version with some telemetry like storing PC-CPU states to disk, I could also try to run it and send the telemetry back.

THe issue seems to be not available in R917 (and already in R872) anymore  

... might get closed...

THe issue seems to be not available in R917 (and already in R872) anymore V1.0.0 (R853) crash induced by tunnel + logic analyzer condition trigger 1f603

... might get closed...

Great!