Node process die when running in arm64v8 (qemu) docker

edit

Hello,

I'm unable to load any node when running ROS inside a qemu arm64v8 docker. The node process dies shortly after being started. The host computer is running Ubuntu 18.04 64 bits. The problem can be easily reproduced as follow:

# Setup qemu to run arm docker on x86 host
sudo apt-get install qemu binfmt-support qemu-user-static
docker run --rm --privileged multiarch/qemu-user-static --reset -p yes

# Launch arm64v8 container with ros image
docker run -it arm64v8/ros:melodic-ros-base /bin/bash

# Install and run roscpp tutorial
apt-get update && apt-get install ros-melodic-roscpp-tutorials
source /opt/ros/melodic/setup.bash
roslaunch roscpp_tutorials talker_listener.launch

Here is the output:

... logging to /root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/roslaunch-8dd91dac75d2-2141.log
Checking log directory for disk usage. This may take a while.
Press Ctrl-C to interrupt
Done checking log file disk usage. Usage is <1GB.

started roslaunch server http://8dd91dac75d2:37539/

SUMMARY
========

PARAMETERS
 * /rosdistro: melodic
 * /rosversion: 1.14.5

NODES
  /
    listener (roscpp_tutorials/listener)
    talker (roscpp_tutorials/talker)

auto-starting new master
process[master]: started with pid [2166]
ROS_MASTER_URI=http://localhost:11311

setting /run_id to a72e1788-b002-11ea-9127-3f45a6660ee1
process[rosout-1]: started with pid [2183]
started core service [/rosout]
process[listener-2]: started with pid [2188]
[rosout-1] process has died [pid 2183, exit code 127, cmd /opt/ros/melodic/lib/rosout/rosout __name:=rosout __log:=/root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/rosout-1.log].
log file: /root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/rosout-1*.log
[rosout-1] restarting process
process[talker-3]: started with pid [2195]
[ INFO] [1592333610.205317895]: hello world 0
qemu:handle_cpu_signal received signal outside vCPU context @ pc=0x6005fc3a
qemu:handle_cpu_signal received signal outside vCPU context @ pc=0x60001737
[ INFO] [1592333610.301632331]: hello world 1
process[rosout-1]: started with pid [2202]
[ INFO] [1592333610.400877386]: hello world 2
[listener-2] process has died [pid 2188, exit code 127, cmd /opt/ros/melodic/lib/roscpp_tutorials/listener __name:=listener __log:=/root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/listener-2.log].
log file: /root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/listener-2*.log
[ INFO] [1592333610.500888051]: hello world 3
[ INFO] [1592333610.600877653]: hello world 4
qemu:handle_cpu_signal received signal outside vCPU context @ pc=0x6005fc3a
qemu:handle_cpu_signal received signal outside vCPU context @ pc=0x60001737
[talker-3] process has died [pid 2195, exit code 127, cmd /opt/ros/melodic/lib/roscpp_tutorials/talker __name:=talker __log:=/root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/talker-3.log].
log file: /root/.ros/log/a72e1788-b002-11ea-9127-3f45a6660ee1/talker-3*.log

The crash seems to be related to these messages:

qemu:handle_cpu_signal received signal outside vCPU context @ pc=0x6005fc3a

and I didn't find any relevant solution on Google. Any idea where to start? Should I submit this to the bug tracker?