Add comma body firmware by jhuang2006 · Pull Request #2291 · commaai/panda

jhuang2006 · 2025-10-15T21:30:34Z

No description provided.

adeebshihadeh · 2025-10-15T21:56:06Z

board/boards/body.h

you shouldn't need this one. this is for pandas

adeebshihadeh · 2025-10-15T21:56:13Z

board/boards/board_declarations.h

same with this file

adeebshihadeh · 2025-10-15T21:56:43Z

board/body/boards/motor_control.h

+// M1 drive: PB8 -> TIM16_CH1, PB9 -> TIM17_CH1, PE2/PE3 enables
+// M2 drive: PA2 -> TIM15_CH1, PA3 -> TIM15_CH2, PE8/PE9 enables
+
+#ifndef MOTOR_PWM_TIMER_CLOCK_HZ


what's up with all the ifndef??

adeebshihadeh · 2025-10-15T22:25:58Z

board/body/boards/motor_control.h

+#define MOTOR_SPEED_CONTROL_UPDATE_PERIOD_US 1000U
+#endif
+
+#define MOTOR_CONTROL_UNUSED __attribute__((unused))


what's this for?

adeebshihadeh

Let's clean this up and get it merged as clean base FW that just runs on the body board, then we can add the rest of the functionality in smaller followup PRs.

adeebshihadeh · 2025-10-18T16:59:26Z

board/body/boards/board_body.h

+void board_body_init_bootloader(void) {
+}
+
+void board_body_tick(void) {
+}


remove all these things that are unused. the board struct isn't the same as panda's so you don't have to populate all these unused fields

adeebshihadeh · 2025-10-18T16:59:38Z

board/body/boards/board_body.h

+  // Body board: PD0 (RX), PD1 (TX)
+  set_gpio_pullup(GPIOD, 0, PULL_NONE);
+  set_gpio_alternate(GPIOD, 0, GPIO_AF9_FDCAN1);


wrong indentation

adeebshihadeh · 2025-10-18T16:59:52Z

board/body/boards/board_body.h

+uint32_t board_body_read_voltage_mV(void) {
+  return 0U;
+}
+
+uint32_t board_body_read_current_mA(void) {
+  return 0U;
+}


remove these too

adeebshihadeh · 2025-10-18T17:00:51Z

board/body/boards/board_declarations.h

+
+// ******************* Definitions ********************
+#define HW_TYPE_UNKNOWN 0U
+#define HW_TYPE_BODY 0xB0U


i like the b prefix. let's call it 0xb1 or 0xb2 though?

adeebshihadeh · 2025-10-18T17:02:21Z

board/body/main_comms.h

we should pull out all these common ones between panda, jungle, and body. but we can do that in a followup

adeebshihadeh · 2025-10-18T17:02:48Z

board/stm32h7/peripherals.h

+  // Body board: PD0 (RX), PD1 (TX)
+  set_gpio_pullup(GPIOD, 0, PULL_NONE);
+  set_gpio_alternate(GPIOD, 0, GPIO_AF9_FDCAN1);
+
+  set_gpio_pullup(GPIOD, 1, PULL_NONE);
+  set_gpio_alternate(GPIOD, 1, GPIO_AF9_FDCAN1);


can you do this in the board init?

adeebshihadeh · 2025-10-21T20:01:34Z

board/body/boards/board_body.h

+void board_body_set_can_mode(uint8_t mode) {
+  if (mode == CAN_MODE_NORMAL) {


doesn't need (or have) modes

adeebshihadeh · 2025-10-21T20:03:12Z

board/body/__init__.py

+  # ****************** Motor Control *****************
+
+  def motor_set_speed(self, motor: int, speed: int) -> None:
+    assert motor in (1, 2), "motor must be 1 or 2"


we should make nice enums for the motors. like which one is 1 and 2?

adeebshihadeh · 2025-10-21T20:04:02Z

board/body/boards/motor_encoder.h

this should go up one level. the boards/ folder is for different body motherboards

adeebshihadeh · 2025-10-21T20:04:09Z

board/body/boards/motor_control.h

adeebshihadeh · 2025-10-21T20:04:34Z

board/body/boards/board_declarations.h

+  const uint8_t led_pin[3];
+  const uint8_t led_pwm_channels[3]; // leave at 0 to disable PWM
+  board_init init;
+  board_init_bootloader init_bootloader;


this is unused, along with a bunch of other stuff

compile errors if we remove init_bootloader and led driver requires pwm channels to be declared. removed all other unused

board/drivers/pwm.h

adeebshihadeh · 2025-10-21T20:23:48Z

@rexblade21 This is pretty close, but there's still a decent amount to trim here. Before requesting another review, I recommend talking to an LLM about the PR, e.g. "Do you see any way we can simplify parts of this diff?"

board/body/boards/board_body.h

adeebshihadeh · 2025-10-22T21:58:57Z

board/body/boards/board_body.h

+  .led_GPIO = {GPIOC, GPIOC, GPIOC},
+  .led_pin = {7, 7, 7},


this is confusing but i assume done to maintain compatibility with panda? if so, leave a comment

also, do we only have a single color LED?

yes to maintain compatibility with led driver. and yes
given that its just a single blinking heartbeat led, we don't really have to use the led driver?

adeebshihadeh · 2025-10-22T21:59:04Z

board/body/boards/board_body.h

+  .led_pin = {7, 7, 7},
+  .init = board_body_init,
+  .enable_can_transceiver = board_body_enable_can_transceiver,
+  .has_spi = false,


don't need this

adeebshihadeh · 2025-10-22T21:59:23Z

board/body/boards/board_declarations.h

+
+// ******************** Prototypes ********************
+typedef void (*board_init)(void);
+typedef void (*board_init_bootloader)(void);


board/body/main_comms.h

board/body/boards/board_declarations.h

adeebshihadeh · 2025-10-22T22:05:33Z

board/body/stm32h7/board.h

./board/early_init.h: In function 'early_initialization':
./board/early_init.h:56:3: error: implicit declaration of function 'detect_board_type' [-Werror=implicit-function-declaration]
56 | detect_board_type();

seems to be required for early_init. moving to board_body.h

adeebshihadeh · 2025-10-22T22:06:47Z

board/body/can.h

+  if (!body_can_bus_ready(bus)) {
+    return false;
+  }


does this matter? don't we just recover anyway? or is this to prevent bus off such that we don't need to recover?

adeebshihadeh · 2025-10-24T15:53:03Z

board/drivers/fdcan.h


    safety_rx_invalid += safety_rx_hook(&to_push) ? 0U : 1U;
+#ifdef PANDA_BODY
+    body_can_safety_rx(&to_push);


this should just be safety_rx_hook

adeebshihadeh · 2025-10-24T20:14:56Z

You need to ignore the body for the MISRA mutation tests: https://github.com/commaai/panda/actions/runs/18775663943/job/53570424244?pr=2291

* motors * can * cleanup unused stuff * initial clean * more clean * remove integral and derivative clamps, revert pwm driver to original * remove integral and derivative clamps, revert pwm driver to original * remove integral and derivative clamps, revert pwm driver to original * dont need this for now * clean * fix can rx and can version error * ignore body for misra mutation test * fix bus recovery, remove body rx hook

* spi debug (commaai#2292) Co-authored-by: Comma Device <device@comma.ai> * Add comma body firmware (commaai#2291) * motors * can * cleanup unused stuff * initial clean * more clean * remove integral and derivative clamps, revert pwm driver to original * remove integral and derivative clamps, revert pwm driver to original * remove integral and derivative clamps, revert pwm driver to original * dont need this for now * clean * fix can rx and can version error * ignore body for misra mutation test * fix bus recovery, remove body rx hook * SocketPanda improvements (commaai#2297) * SocketPanda improvements * implement timeouts * Fix mcu_type for deprecated pandas (commaai#2296) * skip 1024 samples to settle, around 22ms (commaai#2295) * skip 1024 samples to settle, around 22ms * smaller diff * Fix mcu_type in jungle (commaai#2300) should fix jungle * Cuatro siren (commaai#2294) * Fix siren * fix MISRA * remove MISRA suppression * disable amp * fix rebase mistake * fix fault * diff audio * misra --------- Co-authored-by: Comma Device <device@comma.ai> * add double buffer for microphone (commaai#2299) * add double buffer for microphone * comment * fix size check * Revert `mcu_type` changes (commaai#2303) * Revert "Fix mcu_type for deprecated pandas (commaai#2296)" This reverts commit 6c9064c. * Revert "Fix mcu_type in jungle (commaai#2300)" This reverts commit 1e8fa51. * it's just unsupported * cleanup fan scripts * garbage collect always-true condition check (commaai#2305) garbage collect dead code * CI: use tags for cppcheck update cppcheck doesn't always create a release for each tag * lil more * Adjust `gitversion` handling to include null terminator in length calculations. (commaai#2309) Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com> * fix up fan HITL test (commaai#2317) * fix up fan HITL test * cleanup * times two * even simpler * append * one more... * speed it up now * oops * Build everything before jungle recover (commaai#2316) * make sure everything is built, including the bootloader * also for flash * Align delay and compensate (commaai#2318) * align delay and compensate * empty * rm -rf drivers/spi/; kernel driver isn't needed * [bot] Update cppcheck to 2.19.1 (commaai#2254) * [bot] Update cppcheck to 2.19.1 * update coverage table --------- Co-authored-by: Vehicle Researcher <user@comma.ai> Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com> * Revert "[bot] Update cppcheck to 2.19.1 (commaai#2254)" This reverts commit ae769db. * remove mcu_type (commaai#2324) * remove mcu_type from Panda, it's redundant since all supported devices are H7 * revert disable automatic CAN-FD switching H7 check * revert original can no longer be flashed error * assert, assert, assert * normal reset is fine * fix usb connect on macos (commaai#2326) fix claim on macos * Revert "remove mcu_type (commaai#2324)" This reverts commit 8922b48. * remove mcu_type (commaai#2327) * remove mcu_type from Panda, it's redundant since all supported devices are H7 * revert disable automatic CAN-FD switching H7 check * revert original can no longer be flashed error * assert, assert, assert * normal reset is fine * remove assert from recover * update pyproject.toml: include panda.python and panda.board (commaai#2328) * improve HITL robustness (commaai#2333) * windows: fix fcntl import (commaai#2329) * windows: fix fcntl import * fix indentation * add windows ci * make sure CI catches import error * Revert "make sure CI catches import error" This reverts commit b18043a. --------- Co-authored-by: Adeeb Shihadeh <adeebshihadeh@gmail.com> Co-authored-by: Comma Device <device@comma.ai> Co-authored-by: Jason Huang <91160588+rexblade21@users.noreply.github.com> Co-authored-by: Willem Melching <willem.melching@gmail.com> Co-authored-by: Igor Biletski <briskspirit@users.noreply.github.com> Co-authored-by: Robbe Derks <robbe.derks@gmail.com> Co-authored-by: Jason Young <46612682+jyoung8607@users.noreply.github.com> Co-authored-by: downquark7 <mail2emn@gmail.com> Co-authored-by: commaci-public <60409688+commaci-public@users.noreply.github.com> Co-authored-by: Vehicle Researcher <user@comma.ai> Co-authored-by: Andi Radulescu <andi.radulescu@gmail.com>

jhuang2006 force-pushed the body branch 8 times, most recently from 6aec3bf to 47caf95 Compare October 15, 2025 21:51

motors

07b26ee

jhuang2006 force-pushed the body branch from 47caf95 to 07b26ee Compare October 15, 2025 21:53

adeebshihadeh reviewed Oct 15, 2025

View reviewed changes

jhuang2006 marked this pull request as draft October 16, 2025 01:37

can

b1c2db2

adeebshihadeh reviewed Oct 18, 2025

View reviewed changes

cleanup unused stuff

5031e10

jhuang2006 marked this pull request as ready for review October 21, 2025 08:19

jhuang2006 requested a review from adeebshihadeh October 21, 2025 08:20

adeebshihadeh reviewed Oct 21, 2025

View reviewed changes

jhuang2006 added 6 commits October 21, 2025 16:13

initial clean

9938bc9

more clean

a3719da

remove integral and derivative clamps, revert pwm driver to original

ee7b29b

remove integral and derivative clamps, revert pwm driver to original

9a1393c

remove integral and derivative clamps, revert pwm driver to original

4e7c576

dont need this for now

ec4d387

jhuang2006 requested a review from adeebshihadeh October 22, 2025 10:06

adeebshihadeh requested changes Oct 22, 2025

View reviewed changes

jhuang2006 added 2 commits October 22, 2025 23:50

clean

0962b92

fix can rx and can version error

1467048

adeebshihadeh reviewed Oct 24, 2025

View reviewed changes

jhuang2006 added 2 commits October 24, 2025 16:10

ignore body for misra mutation test

9054960

fix bus recovery, remove body rx hook

cf7e4ff

adeebshihadeh changed the title ~~Body Refactor~~ Add comma body firmware Oct 25, 2025

adeebshihadeh merged commit 515ac45 into master Oct 25, 2025
6 checks passed

adeebshihadeh deleted the body branch October 25, 2025 00:14

mmoo758 added a commit to mmoo758/panda that referenced this pull request Feb 24, 2026

Add comma body firmware (commaai#2291)

2f29e38

		void board_body_set_can_mode(uint8_t mode) {
		if (mode == CAN_MODE_NORMAL) {

Conversation

jhuang2006 commented Oct 15, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

adeebshihadeh left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

adeebshihadeh commented Oct 21, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jhuang2006 Oct 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

adeebshihadeh commented Oct 24, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

jhuang2006 Oct 22, 2025 •

edited

Loading