Oniro Zephyr-blinky image doesn't work on Arduino board

The module affected

Arduino board

What is the action to reproduce the bug?

Build and flash Arduino board with Zephyr-blinky ikage

What is the result you see?

Nothig

What is the result you expect?

A led blink on Arduino board

Frequency?

Always

Tested version

Oniro Zephyr-blinky

Do you know any workaround of this issue?

Do you have a fix for this issue?

assigned to @landgraf

moved from oniro#585 (moved)

changed milestone to %goofy alpha

added IssueBug teammaintenance labels

assigned to @stefanschmidt

assigned to @landgraf and unassigned @stefanschmidt

changed milestone to %goofy beta

added 1 deleted label

changed title from Oniro Zephyr-blinky image deosn't work on Arduino board to Oniro Zephyr-blinky image doesn't work on Arduino board

added SeverityCritical label

marked this issue as related to oniro#632 (closed)

marked this issue as related to oniro#633 (closed)

removed the relation with oniro#633 (closed)

@lucazizolfi can you connect to the board with screen/minicom and check if it starts blinking?

@landgraf I flashed the board with blinky's image and now it works properly, the LED flashes

This causes by ea7cea6c and following df59d827

@brgl @agherzan I think we should change the behaviour of the switch. CI should add ONIRO_ENABLE_ACM0 or similar not users add ONIRO_NO_ACM0 to make blinky working again

@lucafavaretto workaround is as simple as: echo ONIRO_NO_ACM0 = "1" >> conf/local.conf

Lowering severity

@landgraf if the workaround works (we gonna test it today) the issue will change iteration to July or will be fixed in the current iteration anyway?

@landgraf the workaround indicated above (the change in the conf/localconf file) is to be done before or after launching the build?

either before or ONIRO_NO_ACM0 = "1" bitbake ....

@landgraf I agree with this change. With the mention that it should be documented.

added SeverityNormal label and removed SeverityCritical label

@lucafavaretto This is not a regression though - it used to work like that ever since I started working on oniro zephyr. The app waits for the ACM0 to be opened on the host. The QA team asked me explicitly to keep it this way.

@brgl can you elaborate please? which QA team? Both Oniro and Zephyr have in their documentation something like "After flashing, the LED starts to blink. Blinky does not print to the console.". from user's point of view this "feature" is completely broken. As I mentioned this logic should be reverted to "blink by default", wait for connection if flag has been specified

@landgraf if you take current dunfell oniro and build the blinky image, it won't start blinking until you start a console on the ACM0 device. It was Chase Qi who requested it be kept the same in kirkstone IIRC. I don't often look at the docs but maybe they're wrong. I agree it's a broken idea hence this variable.

I don't protect it's a regression (I'm pretty sure it's not) but that should be fixed anyway

If you agree - the fix is super easy - just reverse the logic of that variable. Have ONIRO_USE_ACM0 with a positive logic instead.

I do. This was my initial suggestion. We should communicate the change to all interested teams though

mentioned in commit landgraf/meta-zephyr@5b34462e

mentioned in merge request !12 (merged)

@landgraf would you please also try the test images? Here is the guide to build zephyr test images https://gitlab.eclipse.org/eclipse/oniro-core/meta-zephyr/-/blob/oniro/kirkstone/README.txt. After test image flashed, please check if:

• acm0 device appears
• test output observed with acm0

@chaseqi Can you help me with validation? Here what I have now:

Qemu-x86

MACHINE=qemu-x86 bitbake zephyr-kernel-test-all -c testimage

RESULTS:
RESULTS - zephyr.ZephyrTest.test_boot_zephyr: PASSED (6.10s)
SUMMARY:
context () - Ran 1 test in 6.105s
context - OK - All required tests passed (successes=1, skipped=0, failures=0, errors=0)
Traceback (most recent call last):
  File "/mnt/builds/oniroproject/goofy/sources/oe-core/meta/lib/oeqa/core/decorator/__init__.py", line 35, in wrapped_f
    return func(*args, **kwargs)
  File "/mnt/builds/oniroproject/goofy/sources/meta-zephyr/meta-zephyr-core/lib/oeqa/runtime/cases/zephyr.py", line 19, in test_boot_zephyr
    line = self.target.serial_readline().decode("utf-8")
  File "/mnt/builds/oniroproject/goofy/sources/oe-core/../meta-zephyr/meta-zephyr-core/lib/oeqa/controllers/zephyrtargetcontrol.py", line 62, in serial_readline
    return self.runner.serial_readline()
  File "/mnt/builds/oniroproject/goofy/sources/oe-core/../meta-zephyr/meta-zephyr-core/lib/oeqa/utils/qemuzephyrrunner.py", line 225, in serial_readline
    self.log(line)
  File "/mnt/builds/oniroproject/goofy/sources/oe-core/meta/lib/oeqa/utils/qemurunner.py", line 118, in log
    self.msg += msg
  File "/mnt/builds/oniroproject/goofy/sources/oe-core/meta/lib/oeqa/core/decorator/oetimeout.py", line 18, in _timeoutHandler
    raise OEQATimeoutError("Timed out after %s "
oeqa.core.exception.OEQATimeoutError: Timed out after 120 seconds of execution

failed in this way with and without the change. Is it supposed to work?

Arduino:

MACHINE=arduino-nano-33-ble bitbake -c flash_usb zephyr-kernel-test-early_sleep MACHINE=arduino-nano-33-ble bitbake -c flash_usb zephyr-kernel-test-xip

Works. /dev/ttyACM0 exists

Hi @landgraf I haven't tried zephyr testing with OEQA framework on qemu. We plan to run testing on the physical board directly. I think the trace shows that there is no output on the seiral console after 2 minutes wait.

Thanks for the testing on the board. I will create an issue to enable testing on the nano board in CI. CC @sradakovi @aszpilko

Hi @landgraf I tested with you branch on my nano board. The early-sleep.bin can be flashed, and /dev/ttyACM0 does exist after flashing. This is good.

However, even after power off/on, I don't see any output on the /dev/ttyACM0 port. I am not sure if it is just me. Would you please check if you see any output on you board when you have time?

Here is the steps I used.

1. Press the reset button twice to enter flashing mode.
2. Flash the test image with command: ./bossac --port=/dev/ttyACM0 -d -U -i -e -w early-sleep.bin. Log: https://paste.debian.net/1245706/
3. Check test output with minicom(115200 8N1) on the /dev/ttyACM0 port. No output observed.

For zephyr testing, we need to collect test output from the serial port to determine it is passing/failing. Make sense? This request probably should be covered in a different issue. It has been discussed before.

CC @aszpilko

@chaseqi can you please confirm you can see the expected output with current kirkstone (without my changes). It doesn't work for me.

If you don't - it's different issue and should be reported

@landgraf This has been reported on kirkstone before. I remember @brgl figured it is toolchain related. It works on dunfell, but doesn't work on kirkstone.

@chaseqi right. We cannot fix all possible issue with one change. The MR fixes blinky usecases and (hopefully) doesn't introduce regression in testing. Please file new issue to track the missed output bug. It may have same root cause as oniro#584 (moved)

Agree. Lets track it with the oniro#584 (moved). I think it is the same issue.

mentioned in commit landgraf/meta-zephyr@ad313427

mentioned in issue oniro#584 (moved)

mentioned in commit landgraf/meta-zephyr@b480047e

marked this issue as related to oniro#675 (closed)

mentioned in commit landgraf/meta-zephyr@7da6af58

mentioned in commit landgraf/meta-zephyr@8d775ed4

closed with merge request !12 (merged)

added ValidationReady label

marked this issue as related to oniro#678 (closed)

added ValidationPassed label and removed ValidationReady label