flatsat & canbus: Create flatsat app with CANBus implementation by TheArchons · Pull Request #147 · utat-ss/finch-firmware

TheArchons · 2025-12-20T20:27:46Z

main.c contains code of a working CANBus implmenentation that sends/recieves messages to/from the RaspberryPi on the flatsat.

The main issue was the device tree, which I needed to update to support CAN. ~~The external clock (hse) didn't work so I needed to use the internal one (hsi). I also changed pll to use hsi instead of hse.~~ hse works, the issue was that you needed to manually specify which clocks to use.

mshinjo

Just a general comment and I haven’t fully reviewed the changes yet, but:

Can we create separate patches for different areas? For example, hardware definitions (devicetree) should be in their own patch.
Can we update the CI workflow to build the new app as well if we are adding a new app
Is apps/can-bus/prj.conf intentional? What is its purpose? Is it meant to be used with -DCONF?

TheArchons · 2025-12-20T21:14:25Z

Can we create separate patches for different areas? For example, hardware definitions (devicetree) should be in their own patch.

Just a general comment and I haven’t fully reviewed the changes yet, but:
1. Can we create separate patches for different areas? For example, hardware definitions (devicetree) should be in their own patch.

You're right, I'll split this into two PRs

2. Can we update the CI workflow to build the new app as well if we are adding a new app

Sure

3. Is apps/can-bus/prj.conf intentional? What is its purpose? Is it meant to be used with -DCONF?

Unintentional, fixed

mshinjo · 2025-12-20T21:27:51Z

You're right, I'll split this into two PRs

It doesn't have to be a separate PR, but a separate patch (commit)
This should PR should contain three patches: one to modify dts, one to add a new app, and one to modify CI

mshinjo · 2025-12-20T21:29:09Z

The main issue was the device tree, which I needed to update to support CAN. The external clock (hse) didn't work so I needed to use the internal one (hsi). I also changed pll to use hsi instead of hse.

Do you know if it booted at all?

TheArchons · 2025-12-20T21:33:54Z

The main issue was the device tree, which I needed to update to support CAN. The external clock (hse) didn't work so I needed to use the internal one (hsi). I also changed pll to use hsi instead of hse.

Do you know if it booted at all?

What do you mean by it? can_start didn't have any issues with hse, but the timings must have been off or something since it was never received by the RPI. Might be worth looking into why, but I couldn't figure it out.

chopikus · 2026-01-18T02:35:33Z

So @TheArchons are you using HSE or HSI? The last commit suggests you are still using HSI. Also the copilot wants to change it to HSE? Can you update the commit message that HSE works? Feel free to add more info why did you think it didn't work (maybe it was a different configuration). This would help a lot to someone who will work on it later

chopikus · 2026-01-18T02:35:50Z

Otherwise I'm glad it works 🔥

boards/obc/obc.dts

x4132 · 2026-01-18T02:48:37Z

lgtm

also imagine force pushing, skill issue

TheArchons · 2026-01-18T02:56:10Z

So @TheArchons are you using HSE or HSI? The last commit suggests you are still using HSI. Also the copilot wants to change it to HSE? Can you update the commit message that HSE works? Feel free to add more info why did you think it didn't work (maybe it was a different configuration). This would help a lot to someone who will work on it later

Oh yea I forgot to change the commit messages, thanks.

This provides the configuration needed for CANBus to work on obc. Signed-off-by: Alexander Li <github@thearchons.xyz>

This is mostly the same as the OBC dts in #147, but for PAY. Signed-off-by: Alexander Li <github@thearchons.xyz>

mshinjo

Have we decided between using a separate flatsat image versus developing in the flight software app?

apps/flatsat/src/main.c

This is mostly the same as the OBC dts in #147, but for PAY. Signed-off-by: Alexander Li <github@thearchons.xyz>

I am not sure why this build didn't fail before, but we need to define the CAN configuration for the g431rb. Signed-off-by: Alexander Li <github@thearchons.xyz>

apps/obc/src/main.c

apps/obc/CMakeLists.txt

apps/obc/Kconfig

mshinjo · 2026-03-12T01:16:19Z

apps/obc/src/main.c

+LOG_MODULE_REGISTER(flatsat);
+
+const struct device *const can = DEVICE_DT_GET(DT_CHOSEN(zephyr_canbus));
+
+void pi_rx_callback(const struct device *dev, struct can_frame *frame, void *user_data)
+{
+	LOG_INF("RECEIVED CAN MESSAGE WITH ID [%x]", frame->id);
+
+	const int dlc = frame->dlc;
+
+	for (int i = 0; i < dlc; i++) {
+		LOG_INF("%x", frame->data[i]);
+	}
+}

 int main(void)
 {
-	while (1) {
-		LOG_INF("obc");
-		k_msleep(1000);
+	int ret;
+
+	LOG_INF("Starting CAN send");
+
+	LOG_INF("DEVICE IS READY: %d", device_is_ready(can));
+
+	if (IS_ENABLED(CONFIG_CAN_MODE_LOOPBACK)) {
+		// use loopback mode (send to itself) for debugging
+		ret = can_set_mode(can, CAN_MODE_LOOPBACK);
+		if (ret != 0) {
+			LOG_ERR("Error setting loopback mode [%d]", ret);
+			return 0;
+		}
+		LOG_INF("Loopback mode enabled");
 	}
+
+	ret = can_start(can);
+
+	LOG_INF("CAN STARTED: %d", ret);
+
+	// what to filter for regarding receiving messages
+	const struct can_filter pi_filter = {
+		.flags = 0U,
+		.id = 0x100, // when receiving messages from the pi this must match (e.g. to send 100#ABCD .id needs to be 0x100)
+		.mask = CAN_STD_ID_MASK
+	};
+
+	can_add_rx_filter(can, &pi_rx_callback, NULL, &pi_filter);
+
+	while (true) {
+		struct can_frame frame = {
+			.flags = 0,
+			.id = 0x100,
+			.dlc = 4, // data length. Must match the number of elements in .data
+			.data = { 0xde, 0xad, 0xbe, 0xef }
+		};
+		LOG_INF("Waiting 5 seconds");
+		k_msleep(5000);
+		ret = can_send(can, &frame, K_MSEC(3000), NULL, NULL);
+		if (ret) {
+			LOG_ERR("Sending failed [%d]", ret);
+			continue;
+		}
+
+		LOG_INF("CAN sent successfully");
+	}
+


This isn't OBC specific, so I would say that we clean this up and put it under lib/, and create a thread in apps.

In apps, only thread creations and app specific init code should live.
So in apps, in terms of number of lines, it should be short. My guess is that it would be max 200 lines.

Also, we really should start using libcsp. The CAN transportation shouldn't be happening at the Zephyr API layer. Otherwise, we will starting to create technical debt... and it would be hard to fix and clean up... (which happened before the Zephyr introduction)

@alexapostolu FYI I have some libcsp stuff, which should be PR ready this weekend.

@mshinjo Are you going to be in the lab on Saturday? We can talk about this more and get the libcsp stuff merged too

@alexapostolu 50/50 not sure yet. I can probably visit around 5PM. Talking in-person would be way easier and faster.

main.c contains a working CANBus implementation that sends and receives messages to/from the Raspberry Pi on the flatsat. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons requested review from Kevin-Caldwell, alexapostolu, chopikus, coby1kenobi, mshinjo and swarnavaghosh04 as code owners December 20, 2025 20:27

TheArchons added the DNM DO NOT MERGE (ONLY THOSE WHO PUT THIS LABEL CAN REMOVE IT) label Dec 20, 2025

alexapostolu requested review from joy-xu25 and jpmedinagl December 20, 2025 20:30

TheArchons force-pushed the can-bus branch 11 times, most recently from 98229de to d13a7a4 Compare December 20, 2025 21:04

mshinjo reviewed Dec 20, 2025

View reviewed changes

TheArchons force-pushed the can-bus branch 2 times, most recently from 8d54a19 to 6f5c569 Compare December 20, 2025 21:13

TheArchons force-pushed the can-bus branch 3 times, most recently from 6fc4665 to 9e69dad Compare December 20, 2025 21:26

chopikus requested changes Jan 18, 2026

View reviewed changes

boards/obc/obc.dts Outdated Show resolved Hide resolved

TheArchons force-pushed the can-bus branch from 9e3b366 to 77cdb0b Compare January 18, 2026 02:48

TheArchons force-pushed the can-bus branch from 77cdb0b to 6504cb9 Compare January 18, 2026 02:56

obc: dts: canbus: Update devicetree to enable CANBus support for obc

6fc5e1e

This provides the configuration needed for CANBus to work on obc. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons force-pushed the can-bus branch from 6504cb9 to 16f4506 Compare January 24, 2026 20:01

TheArchons requested a review from chopikus January 24, 2026 20:41

mshinjo mentioned this pull request Feb 1, 2026

obc: add framework for modes of operation #152

Draft

alexapostolu previously approved these changes Feb 8, 2026

View reviewed changes

alexapostolu removed the DNM DO NOT MERGE (ONLY THOSE WHO PUT THIS LABEL CAN REMOVE IT) label Feb 8, 2026

TheArchons added a commit that referenced this pull request Feb 8, 2026

boards: pay: pay.dts: Add CAN configuration to pay.dts

f013c23

This is mostly the same as the OBC dts in #147, but for PAY. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons mentioned this pull request Feb 8, 2026

boards: pay: pay.dts: Add CAN configuration to pay.dts #155

Merged

mshinjo reviewed Feb 8, 2026

View reviewed changes

apps/flatsat/src/main.c Outdated Show resolved Hide resolved

alexapostolu pushed a commit that referenced this pull request Feb 8, 2026

boards: pay: pay.dts: Add CAN configuration to pay.dts

53ccd21

This is mostly the same as the OBC dts in #147, but for PAY. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons dismissed alexapostolu’s stale review via 79547c4 March 11, 2026 23:14

TheArchons force-pushed the can-bus branch 5 times, most recently from 40c2a2a to 2bc1fe7 Compare March 12, 2026 00:25

boards: overlays: g431rb: Add CAN configuration for the Nucleo g431rb

36d95c8

I am not sure why this build didn't fail before, but we need to define the CAN configuration for the g431rb. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons force-pushed the can-bus branch from 2bc1fe7 to d252570 Compare March 12, 2026 00:27

TheArchons requested review from alexapostolu and mshinjo March 12, 2026 00:32

mshinjo reviewed Mar 12, 2026

View reviewed changes

obc: Add CANBus implementation to the obc app

0c9cf61

main.c contains a working CANBus implementation that sends and receives messages to/from the Raspberry Pi on the flatsat. Signed-off-by: Alexander Li <github@thearchons.xyz>

TheArchons force-pushed the can-bus branch from d252570 to 0c9cf61 Compare March 14, 2026 18:52

Conversation

TheArchons commented Dec 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mshinjo left a comment

Choose a reason for hiding this comment

Uh oh!

TheArchons commented Dec 20, 2025

Uh oh!

mshinjo commented Dec 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mshinjo commented Dec 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TheArchons commented Dec 20, 2025

Uh oh!

chopikus commented Jan 18, 2026

Uh oh!

chopikus commented Jan 18, 2026

Uh oh!

Uh oh!

x4132 commented Jan 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TheArchons commented Jan 18, 2026

Uh oh!

mshinjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mshinjo Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

alexapostolu Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

mshinjo Mar 12, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

8 participants

TheArchons commented Dec 20, 2025 •

edited

Loading

mshinjo commented Dec 20, 2025 •

edited

Loading

mshinjo commented Dec 20, 2025 •

edited

Loading

x4132 commented Jan 18, 2026 •

edited

Loading