misc issues with tests/trickle

[ZetaR60]

Description

While running tests/trickle for #8904 on mega-xplained, I have run into a number of different issues. One issue is that prev_now is not being initialized properly in main (it assumes xtimer_now_usec is zero when main starts, which is not the case). This is easily fixed by adding prev_now = xtimer_now_usec(); to the beginning of main. There are also a number of problems where it hangs at various points.

Steps to reproduce the issue

Run tests/trickle with the following variations:

• trunk on mega-xplained
• cpu/atmega_common: return to non-interrupt context swaps #8904 applied to trunk on mega-xplained
• trunk on mega-xplained with prev_now = xtimer_now_usec();
• cpu/atmega_common: return to non-interrupt context swaps #8904 applied to trunk on mega-xplained with prev_now = xtimer_now_usec();

Expected results

tests/trickle runs, printing 5 messages after [START], and then 7 messages after [TRICKLE_RESET], and then prints [SUCCESS].

Actual results

• Trunk hangs after [TRICKLE_RESET]
• cpu/atmega_common: return to non-interrupt context swaps #8904 applied to trunk declares failure (due to not initializing prev_now properly)
• Trunk with prev_now initialized hangs after three messages (I think this is a cpu/atmega_common: Context switching race condition #8896 related problem, and is fixed by cpu/atmega_common: return to non-interrupt context swaps #8904).
• cpu/atmega_common: return to non-interrupt context swaps #8904 applied to trunk and prev_now initialized hangs after [TRICKLE_RESET]

Versions

Board: mega-xplained

Installed compiler toolchains 
-----------------------------
             native gcc: missing
      arm-none-eabi-gcc: missing
                avr-gcc: avr-gcc (Fedora 7.2.0-1.fc27) 7.2.0
       mips-mti-elf-gcc: missing
             msp430-gcc: missing
   riscv-none-embed-gcc: missing
                  clang: missing

Installed compiler libs
-----------------------
   arm-none-eabi-newlib: missing
    mips-mti-elf-newlib: missing
riscv-none-embed-newlib: missing
               avr-libc: "2.0.0" ("20150208")

Installed development tools
---------------------------
                  cmake: cmake version 3.11.0
               cppcheck: missing
                doxygen: missing
                 flake8: missing
                    git: git version 2.14.2
             coccinelle: missing

[cladmi]

Thank for investigating this, I also had problem on tests/trickle with arduino-zero.
Adding the prev_now = xtimer_now_usec(); in main fixes it. I will do a PR for it.

[cladmi]

I implemented your prev_now fix in #9118.

However I misread and thought it also fixed your problem, I can amend my commit message depending on your feedback.

[ZetaR60]

@cladmi I found that prev_now = xtimer_now_usec(); improves the behavior of tests/trickle, but does not solve all of the problems (otherwise I would have just opened a PR). There is still the hang after [TRICKLE_RESET]. Do you get that hang on the platform you are testing (arduino-zero)? I took a look through the trickle code, but I am not sure what is causing it to hang.

[cladmi]

@ZetaR60 I am not seeing any hang on arduino-zero in the trickle test, with the initialization fix, tests are working.
However I am not really using the board for anything, just running the automated tests on IoT-LAB.