Samraksh / eMote

eMote OS -- Multiple Ports (using .NET MF v4.3)
0 stars 0 forks source link

With VS deploy, node starts new program but does not turn on radio #450

Closed WilliamAtSamraksh closed 7 years ago

WilliamAtSamraksh commented 7 years ago

I'm using eMote built from commit e1a8575 since at the moment the reprogrammer conflicts with eMote release 24.

I deploy Fence onto a mote from VS. After deployment the program begins to run. However the radio (RF231) apparently does not turn on. If I use the reset button then all is well.

WilliamAtSamraksh commented 7 years ago

Actually, it's more than that. When I ping the node after VS deploy there's no connection. The program appears to be running however, with the correct value in the LCD displayed.

MichaelAtSamraksh commented 7 years ago

Is the LCD written before the radio is initialized? Is this in release or debug flavor? What is the backtrace if you have a JTAG attached?

WilliamAtSamraksh commented 7 years ago

The problem incidentally is intermittent. Sometimes it's as I described, sometimes after VS deploy everything is good.

Is the LCD written before the radio is initialized?

Yes

Is this in release or debug flavor?

Debug

What is the backtrace if you have a JTAG attached?

Sorry, no JTAG

ChrisAtSamraksh commented 7 years ago

I hope 4be7cd689b7653fff96873954be04a3fafbc13d5 fixes your problem. Try it out.

MichaelAtSamraksh commented 7 years ago

I have observed the radio locking up during soft reboots, like what Visual Studio does when it is finished installing. I bet the radio is remaining locked up as program execution begins. After the soft reboot, the radio lockup is most likely manifesting as a hard fault in netmf_spi.cpp during a call to SPI_I2S_GetFlagStatus(...) surrounded by a TIMEOUT_WAIT macro. The TIMEOUT_WAIT macro triggers the ASSERT macro. In Debug flavor builds, ASSERT will call HAL_AssertEx and cause a hard fault to occur.

... but hopefully Chris's commit that removes the RF231 debug GPIO toggles will fix your problem.

WilliamAtSamraksh commented 7 years ago

@ChrisAtSamraksh: Presently there's a conflict between the reprogrammer HAL code and eMote v. 24. If your commit is based on v. 24 then I'll have to wait until the conflict is sorted.

WilliamAtSamraksh commented 7 years ago

@ChrisAtSamraksh: Commit 4be7cd6 dealt with the RF radio. Is the SI radio ok or does it also need a fix?

ChrisAtSamraksh commented 7 years ago

The SI radio interaction is not changed with that commit. But you mentioned RF231 problem and those GPIO toggles should not be the master branch.

Are you still having rf231 problems or also having SI problems?