opendbc: regen prompt earlier + radar donor pin - #5
Merged
Merged
Conversation
New radars still hold the donor VIN and reject this car. Read F190 when vinValidity sticks, then put that VIN on the GTW feed. Empty VIN keeps current passthrough. Co-authored-by: Cursor <cursoragent@cursor.com>
IRITable never set vinValidity, so auto-read never started. The UI can now request one F190 read. SGUFail was blocking engage on a live uncalibrated radar; HWFail still blocks. Co-authored-by: Cursor <cursoragent@cursor.com>
Empty VIN used to skip 0x560, so panda wrote position from the leftover behind-nosecone flag. Always stream 0x560 so Donor Position is the nibble, including passthrough. Keep a decoded F190 if cleanup times out, and treat a frozen Bosch table plus SGUFail as radarFault. Co-authored-by: Cursor <cursoragent@cursor.com>
VIN character 8 of 2/4 is not dual-motor on 2015 RWD 5YJSA1E2. Forcing Msg2A9_FourWheelDrive on this chassis sprays Bosch tracks in motion. Keep chassis xWD from live 0x398; still impersonate the donor VIN.
NAPRadarIgnoreHwFail left the frozen-table+SGUFail lockout in place, and a compiled params_pyx miss treated ignore as False. Resolve the flag live from nap_conf, Params, then the on-disk file.
IgnoreHwFail was suppressing the frozen-table+SGUFail lockout, so stale tracks could become leadOne and slam aTarget. Keep the toggle for raw HWFail only; frozen && SGUFail still faults after ~5s.
Ignore still covers raw HWFail. Frozen+SGUFail no longer bricks engage when Ignore is on — those ghost tracks are dropped so they cannot become leadOne. Stock-safe radarFault remains when Ignore is off.
… clears xwdValidity
UI overlay toggle. Default off.
Tinkla 0.6.6 did not talk on the radar bus until 0x560 was whole. We were forwarding this-car identity immediately, then switching contracts a second later. Host 0x560 is 10 Hz so the wait is short.
The driver needs more time to add friction brake once regen cannot deliver the request. Fire at a shallower regen command, a smaller shortfall, and 0.5s of evidence instead of 0.8s.
4 tasks
cursor Bot
pushed a commit
to jmbrunick/openpilot
that referenced
this pull request
Sep 16, 2026
Point opendbc_repo at the merge of NotAutopilot/opendbc#5 (3c4da443) so the pedal-chime / earlier-regen-prompt change tracks origin/opendbc nap-dev, not the pre-merge PR head.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
FF nap-dev through the radar-donor-contract SHA openpilot already pins, plus the earlier regen under-delivery prompt (4fd0b31). Needed so the 0.11.1 openpilot submodule pointer exists on origin.