Big ping crashes device and requires restart

[Toad]

In Windows when I run the command:

ping 192.168.0.200 -n 1 -l 65500

this issues a ping of size 65500. After issuing this ping to the Make Controller, the device gets really messed up and requires restarting. I'm using Heavy 1.2 RC4.

[liamMT]

Are you set up to build code for the Controller, or have you just been running the binaries? If you're set up to build the code, I just checked in a couple changes that affect the memory allocation, and I'd be curious to see how they affect your situation.

[Toad]

Replying to liamMT:

Are you set up to build code for the Controller, or have you just been running the binaries? If you're set up to build the code, I just checked in a couple changes that affect the memory allocation, and I'd be curious to see how they affect your situation.

Actually right now, I'm not set up to build code for the controller. I've just been running the binaries. Are you not able to reproduce this?

If those new changes you checked in were part of Heavy 1.2.0 Final, I don't think it worked since I'm still able to reproduce the problem in that version. Here is a video of this happening in 1.2.0 final: http://www.makingthings.com/Members/Toad/ping%20crash.wmv . As you can see in the video from the Make Controller Helper, the make controller is connected via USB and Ethernet. As soon as I issue the standard Windows ping command, you can see the controller dies and disconnects from the computer. The green light on the board stops blinking and the device requires restarting.

I'll see if I can get Crossworks set up on this machine.

[liamMT]

We've seen some similar behavior over TCP, but haven't quite tracked down the problem.

For whatever reason, the Windows ping is not available on my machine currently (must not be on my PATH or something) and other ping programs don't allow such large packets...on OS X, I can send pings with packets as large as it will allow and it seems to work OK. I think these are up to about 5000 bytes or so.

Are you still seeing problems with large OSC/UDP packets?

[liamMT]

This ticket is pretty old - a lot of things have been fixed in the Ethernet system. I'm going to close it for now...please feel free to open a new one if you're still finding you have trouble.

[Toad]

Something has definitely changed from when I originally tested this in 1.2.0 and when I now test it in the latest 1.5.1. I still think there's something wrong. When I do the same thing with 1.5.1 it seems like the board "restarts" and the state of all the digital outputs get reset. This is at least better than what happened before. I don't have any more information about this right now. I'll have to look into this more some other time. I'll leave this closed for now.