tests_hw: Reset USB Port Upon Workflow Start#191
Conversation
mgeeIOL
force-pushed
the
tests/bugfix-fix-cancelled-flash-usb-device-lockup
branch
9 times, most recently
from
ef21fd0 to
3beceff
Compare
Added step to the setup job for the b_u585i_iot02a workflow file to reset the USB connection. This should fix unexpected states the connection is left in as the result of an interrupted west flash command during a cancelled GHA run. Also added additional step to do the same process within the container as it appears the USB connection on the container is affecting the workflow. Signed-off-by: Matthew Gee <mgee@iol.unh.edu>
mgeeIOL
force-pushed
the
tests/bugfix-fix-cancelled-flash-usb-device-lockup
branch
from
3beceff to
72abea0
Compare
|
|
||
| steps: | ||
| - name: Reset USB Connection | ||
| run: usbreset STLINK-V3 # Product name for the b_u585i_iot02a, may have to update if we add additional STM boards to the workflow |
There was a problem hiding this comment.
For clarification this is from libusb as the product name of the board, since this is generic I suspect other STMicroelectronics products may share the same product name seen by usbutils.
Why do you need to reset it twice?
Also, while I don't oppose setting up your container with whatever privileges you want to, giving following full privileges |
Removes the --privileged flag from container usage and instead uses an additional --device to allow the container to manage usb devices as needed. This commit also removes the setup-container step as through testing it is no longer needed. Signed-off-by: Matthew Gee <mgee@iol.unh.edu>
The second reset is from some test runs I did where there was a failure when I only reset it once but succeeded when I did it the way I had it above. I looked into this a bit more and was able to remove it with the additional changes I pushed. Due to this I was also able to remove changes from the docker container. I did some experimenting and was able to reduce the permissions regarding the container and remove the use of This PR should be good to go at this point, let me know if you have any other comments. |
3 participants
Description
When a GHA run was cancelled while on a west flash command the USB port would be left in an unexpected state and future GHA runs would fail as west flash could not successfully run. The aim of this PR is to find a fix
Current planned fix is to add a step to the setup job for the b_u585i_iot02a workflow file to reset the USB connection. This should fix unexpected states the connection is left in as the result of an interrupted west flash command during a cancelled GHA run.