openpilot controlsd: resume from stop doesn't work with stock long

Oct 22 '24 19:10 sshane

Doesn't matter since we're going to use openpilot long

Oct 22 '24 19:10 adeebshihadeh

Right, but this looks like a regression for all the stock long cars (cruiseControl.resume never activates). Will re-visit after Ford.

Oct 22 '24 20:10 sshane

I'm experiencing this also except it's on experimental. I have a '23 Sonata with device on master with ~~commit af73d6084d5ac77935282442c909432eebd676d1 i believe.~~

Edit: The actual commit on the device is 6175106b197f811a0da2dc1856a31abf9f0a9a82

Oct 27 '24 00:10 BBBmau

Can you post a route where this happens?

Oct 28 '24 18:10 sshane

@sshane Edit: correct route is

https://connect.comma.ai/9a31d18cd1a79c69/00000177--035b67fbc9/160/164

9a31d18cd1a79c69/00000177--035b67fbc9/2

Oct 28 '24 19:10 BBBmau

The model in experimental mode failing to resume at traffic lights is different than controlsd setting the carControl.cruiseControl.resume signal based on the planner/lead BTW, that's what this issue is for

Oct 28 '24 20:10 sshane

The model in experimental mode failing to resume at traffic lights is different than controlsd setting the carControl.cruiseControl.resume signal based on the planner/lead BTW, that's what this issue is for

yeah the route is wrong, I'll be providing the route with the actual issue later today. Sorry for the confusion.

Oct 28 '24 21:10 BBBmau

@sshane looking to see what needs to be changed. A bit interesting how on resume it goes to the saved speed but is quickly reverted to be 65

Oct 29 '24 01:10 BBBmau

This works on a recent route, not sure if this was a real issue. Closing

Jan 15 '25 02:01 sshane