drivers: can: stm32fdcan #31061

alexanderwachter · 2020-12-31T12:52:00Z

STM32FDCAN implementation.
This PR implements a generic Bosch M_CAN driver that can be specialized for dedicated SoCs, such as FDCAN (STmicro), Microchip SAM, and NXP.
The specialization for FDCAN (STmicro) is included in this PR.
Tests are added to test the FD mode.

@legoabram has a prototype prototype specialization for SAM0

martinjaeger

Thanks @alexanderwachter for all your effort. I have tested the driver in a classic CAN network already and can confirm that it works very well.

Still need to go through the can_mcan.c, but here are some small comments already.

drivers/can/Kconfig.mcan

boards/arm/nucleo_g474re/nucleo_g474re.dts

drivers/can/Kconfig.stm32fd

include/drivers/can.h

dts/arm/st/g4/stm32g4.dtsi

drivers/can/can_mcan_int.h

drivers/can/can_mcan.c

KozhinovAlexander

Thank you very much for this PR. I would also test with it with my stm32h7 soon and go through ToDo's.

drivers/can/can_mcan.c

drivers/can/can_mcan_int.h

martinjaeger · 2021-01-28T14:30:08Z

dts/arm/st/g4/stm32g4.dtsi

@@ -302,6 +302,54 @@
 			label = "SPI_3";
 		};

+		can {


This leads to a warning during DTS generation:

board_name.dts.pre.tmp:281.7-324.5: Warning (simple_bus_reg): /soc/can: missing or empty reg/ranges property

@mbolivar any suggestions on how we can improve the DT bindings?

Maybe the top-level node could be set to the general CANFD configuration register?

Suggested change

can {

can: can@40006500 {

It possibly misses a reg = <0>;

dts/arm/st/g4/stm32g4.dtsi

KozhinovAlexander · 2021-03-20T15:14:45Z

Exactly. Its for the newer series with FDCAN (M_CAN), like the stm32g4, g0 and H7

In my opinion it is to early to speak about h7. I've tested it on h723 and it needs some changes due to different registers structure. I'll prepare driver update as soon as your development of this branch will be in master.

raul-klg · 2021-03-26T12:52:41Z

Hi! Thanks for your work on this branch.

I think I found a problem running the tests/drivers/can/canfd/ example. I hope this is the right place to discuss this, otherwise please, accept my apologies and let me know where else I should address this information.

I have tried it merged with a slightly more recent zephyr (2235c71) on a nucleo_g0b1re board. For being able to use the driver on the g0b1 board I created a dts overlay for it (it is taken from your stm32g4 proposal):

/ {

    chosen {
                zephyr,can-primary = &can1;
    };

    soc {
                can {
                        compatible = "bosch,m-can-base";
                        #address-cells = <1>;
                        #size-cells = <1>;
                        std-filter-elements = <28>;
                        ext-filter-elements = <8>;
                        rx-fifo0-elements = <3>;
                        rx-fifo1-elements = <3>;
                        rx-buffer-elements = <0>;
                        tx-buffer-elements = <3>;

                        can1: can@40006400 {
                                compatible = "st,stm32-fdcan";
                                #address-cells = <1>;
                                #size-cells = <0>;
                                reg = <0x40006400 0x400>, <0x4000A400 0x350>;
                                reg-names = "m_can", "message_ram";
                                interrupts = <21 0>, <22 0>;
                                interrupt-names = "LINE_0", "LINE_1";
                                status = "disabled";
                                label = "CAN_1";
                        };
                };

    };
};

&can1 {
        pinctrl-0 = <&fdcan1_rx_pa11 &fdcan1_tx_pa12>;
        bus-speed = <125000>;
        sjw = <1>;
        sample-point = <875>;
        bus-speed-data = <1000000>;
        sjw-data = <1>;
        sample-point-data = <875>;
        status = "okay";
};

I then build the tests/drivers/can/canfd/ example and run it on the target with the following serial output:


*** Booting Zephyr OS build zephyr-v2.5.0-1509-gc0835a12fd30  ***
Running test suite canfd_driver
===================================================================
START - test_set_loopback
 PASS - test_set_loopback in 0.1 seconds
===================================================================
START - test_send_receive_std

    Assertion failed at WEST_TOPDIR/sample.git/canfd/src/main.c:117: check_msg: (msg1->id not equal to msg2->id)
ID does not match
 FAIL - test_send_receive_std in 0.13 seconds
===================================================================
START - test_send_receive_fd

Apparently it stalls on the test_send_receive_fd test and also there is an assertion on the previous test_send_receive_std This does not look to me like a successful test run.

For the sake of not cluttering the PR, I am omitting some debug information I could provide if it can be helpful to anyone.

legoabram · 2021-03-26T18:51:54Z

Something I noticed: CAN_MAX_DLC is hard coded to 8 even if FD mode is enabled with a higher CAN_MAX_DLEN. We will likely need an #if tree to make it reflect the max data length. That said, I think changing this define will cause issues since many modules support both classic and CAN-FD, and they will need to use the correct max dlc when checking data depending on the requested mode.

On that same note, a number of the existing drivers depend on this value while iterating through data bytes. Though it will likely work, it's not semantically correct since DLC is no longer explicitly the number of bytes with the introduction of CAN-FD.

alexanderwachter · 2021-03-31T10:42:51Z

@raul-klg do the api tests pass?
I'll rerun the tests on my g4 when time permits.

raul-klg · 2021-03-31T12:15:19Z

Edit:
For the sake of completeness I found out the problem. I reused the register range from stm32g4 and I didn't noticed that the FDCAN message RAM in stm32g0b1 is 0x4000B400 (unlike the stm32g4 0x400A400).

-- Old post
@alexanderwachter Thanks for your reply.
Maybe I am not running the test properly. When I wrote #31061 (comment) I didn't enable the CAN_LOOPBACK and I also didn't connect the nucleo-g0b1re CAN bus to any other device or even bus termination.
I have build now the tests/drivers/can/api application with the same board overlay as detailed in #31061 (comment)

I have run three cases:

default configuration. CONFIG_CAN_LOOPBACK is not set and CONFIG_CAN_STM32FD=y
->Test failed.

*** Booting Zephyr OS build zephyr-v2.5.0-1509-gc0835a12fd30  ***
Running test suite can_driver
===================================================================
START - test_set_loopback
 PASS - test_set_loopback in 0.1 seconds
===================================================================
START - test_send_and_forget
 PASS - test_send_and_forget in 0.1 seconds
===================================================================
START - test_filter_attach
 PASS - test_filter_attach in 0.1 seconds
===================================================================
START - test_receive_timeout
 PASS - test_receive_timeout in 0.101 seconds
===================================================================
START - test_send_callback

    Assertion failed at WEST_TOPDIR/scratch/can-api/src/main.c:609: test_send_callback: (ret not equal to 0)
Missing TX callback
 FAIL - test_send_callback in 0.113 seconds
===================================================================
START - test_send_receive_std

Both CAN loopback and stm32fd driver enabled. CONFIG_CAN_LOOPBACK=y and CONFIG_CAN_STM32FD=y
->Test crashed

I: Init of CAN_LOOPBACK done
*** Booting Zephyr OS build zephyr-v2.5.0-1509-gc0835a12fd30  ***
Running test suite can_driver
===================================================================
START - test_set_loopback
 PASS - test_set_loopback in 0.1 seconds
===================================================================
START - test_send_and_forget
ASSERTION FAIL [z_spin_lock_valid(l)] @ WEST_TOPDIR/zephyr/include/spinlock.h:129
        Recursive spinlock 0x20000d14
E: r0/a1:  0x00000004  r1/a2:  0x00000081  r2/a3:  0x080095cd
E: r3/a4:  0x08005d81 r12/ip:  0x00000002 r14/lr:  0x08005d89
E:  xpsr:  0x21000000
E: Faulting instruction address (r15/pc): 0x0800769a
E: >>> ZEPHYR FATAL ERROR 4: Kernel panic on CPU 0
E: Current thread: (nil) (unknown)
E: Halting system

Only CAN loopback is enabled.can32fd driver is not: CONFIG_CAN_LOOPBACK=y, CONFIG_CAN_STM32FD is not set
->Test passed

I: Init of CAN_LOOPBACK done
*** Booting Zephyr OS build zephyr-v2.5.0-1509-gc0835a12fd30  ***
Running test suite can_driver
===================================================================
START - test_set_loopback
 PASS - test_set_loopback in 0.1 seconds
===================================================================
START - test_send_and_forget
 PASS - test_send_and_forget in 0.1 seconds
===================================================================
START - test_filter_attach
 PASS - test_filter_attach in 0.1 seconds
===================================================================
START - test_receive_timeout
 PASS - test_receive_timeout in 0.101 seconds
===================================================================
START - test_send_callback
 PASS - test_send_callback in 0.1 seconds
===================================================================
START - test_send_receive_std
 PASS - test_send_receive_std in 0.1 seconds
===================================================================
START - test_send_invalid_dlc
E: DLC of 9 exceeds maximum (8)
 PASS - test_send_invalid_dlc in 0.3 seconds
===================================================================
START - test_send_receive_ext
 PASS - test_send_receive_ext in 0.1 seconds
===================================================================
START - test_send_receive_std_masked
 PASS - test_send_receive_std_masked in 0.1 seconds
===================================================================
START - test_send_receive_ext_masked
 PASS - test_send_receive_ext_masked in 0.2 seconds
===================================================================
START - test_send_receive_buffer
 PASS - test_send_receive_buffer in 0.2 seconds
===================================================================
START - test_send_receive_wrong_id
 PASS - test_send_receive_wrong_id in 0.101 seconds
===================================================================
Test suite can_driver succeeded
===================================================================
PROJECT EXECUTION SUCCESSFUL

Please, let me know if I am doing anything wrong and how else could I help.

Urist-McGit · 2021-04-30T12:53:59Z

drivers/can/can_mcan.c

+
+		if (can->ir & CAN_MCAN_IR_ARA) {
+			can->ir  = CAN_MCAN_IR_ARA;
+			LOG_ERR("Acces to reserved address");


I found a small typo:

Suggested change

LOG_ERR("Acces to reserved address");

LOG_ERR("Access to reserved address");

flaminggoat · 2021-05-06T15:35:51Z

I have been testing this CAN implementation on a ATSAME51 using the prototype specialisations mentioned in the OP.
I found their is no register state change isr so I had to add the following to can_sam0.c

static void can_sam0_register_state_change_isr(const struct device *dev,
						can_state_change_isr_t isr)
{
	const struct can_sam0_config *cfg = DEV_CFG(dev);
	const struct can_mcan_config *mcan_cfg = &cfg->mcan_cfg;
	struct can_mcan_reg *can = mcan_cfg->can;
	struct can_mcan_data *mcan_data = &DEV_DATA(dev)->mcan_data;
	struct can_mcan_msg_sram *msg_ram = cfg->msg_sram;

	mcan_data->state_change_isr = isr;

	if (isr == NULL) {
		can->ie &= ~CAN_MCAN_IE_EP;
	} else {
		can->ie |= CAN_MCAN_IE_EP;
	}
}

I also have found that the bitrate does not seem to be working as expected with this device tree I am getting a bitrate of 370k:

&can0 {
	status = "okay";
	bus-speed = <1000000>;
	sjw = <1>;
	sjw-data = <1>;
	sample-point-data = <875>;
	bus-speed-data = <1000000>;
};

alexanderwachter · 2021-05-06T19:28:37Z

@raul-klg the test run on my stm32g474 sucessfully.
west build -b nucleo_g474re zephyr/tests/drivers/can/api and west build -b nucleo_g474re zephyr/tests/drivers/can/canfd

martinjaeger

Still finishing final can_mcan.c review, but here are some remarks already and one point that's not clear for me. Maybe you can comment already, @alexanderwachter?

drivers/can/Kconfig

drivers/can/can_mcan.h

drivers/can/can_mcan.c

drivers/can/can_mcan.h

drivers/can/can_mcan.c

martinjaeger

LGTM, only got above minor comments left.

I've been using this code in classic mode for a while already without any issues.

Thanks again for all the time you put into getting this done @alexanderwachter!

raul-klg · 2021-05-07T12:56:42Z

@raul-klg the test run on my stm32g474 sucessfully.
west build -b nucleo_g474re zephyr/tests/drivers/can/api and west build -b nucleo_g474re zephyr/tests/drivers/can/canfd

Thanks for the verification (and for your additional work). I'll retry when I'm able to. Maybe I need to review how I used the branch.

I believe the issues have been resolved

Add Kconfig option to select appropriate can data field length Signed-off-by: Alexander Wachter <[email protected]>

Implementation of the Bosch M_CAN IP driver. This driver is just the base for a specific SoC implementation. Signed-off-by: Alexander Wachter <[email protected]>

This driver is the SoC specific implementation of the Bosch M_CAN IP. Signed-off-by: Alexander Wachter <[email protected]>

This commit adds the CAN nodes to the STM32g4 SoCs. Signed-off-by: Alexander Wachter <[email protected]>

This commit adds CAN support to the nucleo_g474re board. Signed-off-by: Alexander Wachter <[email protected]>

This commit enhances the CAN API test. In the send_receive tests, we are using two filters and frames at the same time to check if the frame gets dispatched to the correct filter. Signed-off-by: Alexander Wachter <[email protected]>

This commit adds tests for CAN-FD frames. Signed-off-by: Alexander Wachter <[email protected]>

alexanderwachter added platform: STM32 ST Micro STM32 DNM This PR should not be merged (Do Not Merge) area: CAN labels Dec 31, 2020

alexanderwachter requested review from martinjaeger, erwango and ABOSTM December 31, 2020 12:52

alexanderwachter requested a review from FRASTM as a code owner December 31, 2020 12:52

github-actions bot added area: API Changes to public APIs area: Boards area: Devicetree area: Devicetree Binding PR modifies or adds a Device Tree binding labels Dec 31, 2020

alexanderwachter marked this pull request as draft December 31, 2020 12:52

martinjaeger reviewed Jan 1, 2021

View reviewed changes

alexanderwachter force-pushed the can_mcan_stm branch from 6d7e1a9 to af2b381 Compare January 17, 2021 19:07

github-actions bot added the area: Tests Issues related to a particular existing or missing test label Jan 17, 2021

alexanderwachter force-pushed the can_mcan_stm branch from af2b381 to 9198d49 Compare January 17, 2021 19:22

legoabram reviewed Jan 19, 2021

View reviewed changes

drivers/can/can_mcan.c Outdated Show resolved Hide resolved

alexanderwachter mentioned this pull request Jan 24, 2021

Wrong log settings in can_stm32 driver #31573

Closed

KozhinovAlexander suggested changes Jan 24, 2021

View reviewed changes

martinjaeger reviewed Jan 28, 2021

View reviewed changes

alexanderwachter force-pushed the can_mcan_stm branch 3 times, most recently from 107a72f to ecc1e3b Compare January 30, 2021 16:00

martinjaeger reviewed Jan 30, 2021

View reviewed changes

dts/arm/st/g4/stm32g4.dtsi Outdated Show resolved Hide resolved

alexanderwachter force-pushed the can_mcan_stm branch 3 times, most recently from b84aa20 to efdebda Compare January 31, 2021 09:03

alexanderwachter marked this pull request as ready for review January 31, 2021 09:04

alexanderwachter removed the DNM This PR should not be merged (Do Not Merge) label Jan 31, 2021

alexanderwachter added this to the v2.6.0 milestone Jan 31, 2021

alexanderwachter force-pushed the can_mcan_stm branch from bd19adb to fa954ff Compare March 20, 2021 13:35

Urist-McGit reviewed Apr 30, 2021

View reviewed changes

alexanderwachter force-pushed the can_mcan_stm branch 2 times, most recently from b6e1e6a to 2c03505 Compare May 6, 2021 19:24

alexanderwachter requested a review from KozhinovAlexander May 6, 2021 19:25

alexanderwachter force-pushed the can_mcan_stm branch from 2c03505 to 2f3f91a Compare May 6, 2021 21:08

martinjaeger reviewed May 7, 2021

View reviewed changes

drivers/can/Kconfig Outdated Show resolved Hide resolved

drivers/can/can_mcan.h Show resolved Hide resolved

drivers/can/can_mcan.c Show resolved Hide resolved

drivers/can/can_mcan.h Outdated Show resolved Hide resolved

drivers/can/can_mcan.c Outdated Show resolved Hide resolved

martinjaeger approved these changes May 7, 2021

View reviewed changes

erwango approved these changes May 7, 2021

View reviewed changes

str4t0m mentioned this pull request May 7, 2021

Enable CAN peripheral on STM32G0 #34963

Closed

alexanderwachter force-pushed the can_mcan_stm branch from 2f3f91a to 8fc8b15 Compare May 7, 2021 15:50

alexanderwachter added 7 commits May 7, 2021 17:50

drivers: can: Add Kconfig option for CANFD datalenght

d2a6a0a

Add Kconfig option to select appropriate can data field length Signed-off-by: Alexander Wachter <[email protected]>

drivers: can: Implement Bosch M_CAN driver

a1a16b0

Implementation of the Bosch M_CAN IP driver. This driver is just the base for a specific SoC implementation. Signed-off-by: Alexander Wachter <[email protected]>

drivers: can: Implement stm32fd driver

d266ee5

This driver is the SoC specific implementation of the Bosch M_CAN IP. Signed-off-by: Alexander Wachter <[email protected]>

dts: arm: stm32g4: add can bindings to stm32g4 SoCs

b2b5988

This commit adds the CAN nodes to the STM32g4 SoCs. Signed-off-by: Alexander Wachter <[email protected]>

board: nucleo_g474re: Add CAN support.

e937ea2

This commit adds CAN support to the nucleo_g474re board. Signed-off-by: Alexander Wachter <[email protected]>

tests: driver: can: canfd: Add CAN-FD tests

8fc8b15

This commit adds tests for CAN-FD frames. Signed-off-by: Alexander Wachter <[email protected]>

galak merged commit 3b2fd1c into zephyrproject-rtos:master May 7, 2021

JeremyBCD mentioned this pull request May 28, 2021

drivers: can: STM32H7 FDCAN - WIP #35751

Closed

ABOSTM mentioned this pull request May 28, 2021

tests: subsys: canbus: isotp: implementation: fails on nucleo_g474re #35760

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

drivers: can: stm32fdcan #31061

drivers: can: stm32fdcan #31061

alexanderwachter commented Dec 31, 2020 •

edited

Loading

martinjaeger left a comment

KozhinovAlexander left a comment

martinjaeger Jan 28, 2021

alexanderwachter Jan 30, 2021

martinjaeger Jan 30, 2021

erwango May 7, 2021 •

edited

Loading

KozhinovAlexander commented Mar 20, 2021 •

edited

Loading

raul-klg commented Mar 26, 2021

legoabram commented Mar 26, 2021

alexanderwachter commented Mar 31, 2021

raul-klg commented Mar 31, 2021 •

edited

Loading

Urist-McGit Apr 30, 2021

flaminggoat commented May 6, 2021

alexanderwachter commented May 6, 2021 •

edited

Loading

martinjaeger left a comment

martinjaeger left a comment

raul-klg commented May 7, 2021

	LOG_ERR("Acces to reserved address");
	LOG_ERR("Access to reserved address");

drivers: can: stm32fdcan #31061

drivers: can: stm32fdcan #31061

Conversation

alexanderwachter commented Dec 31, 2020 • edited Loading

martinjaeger left a comment

Choose a reason for hiding this comment

KozhinovAlexander left a comment

Choose a reason for hiding this comment

martinjaeger Jan 28, 2021

Choose a reason for hiding this comment

alexanderwachter Jan 30, 2021

Choose a reason for hiding this comment

martinjaeger Jan 30, 2021

Choose a reason for hiding this comment

erwango May 7, 2021 • edited Loading

Choose a reason for hiding this comment

KozhinovAlexander commented Mar 20, 2021 • edited Loading

raul-klg commented Mar 26, 2021

legoabram commented Mar 26, 2021

alexanderwachter commented Mar 31, 2021

raul-klg commented Mar 31, 2021 • edited Loading

Urist-McGit Apr 30, 2021

Choose a reason for hiding this comment

flaminggoat commented May 6, 2021

alexanderwachter commented May 6, 2021 • edited Loading

martinjaeger left a comment

Choose a reason for hiding this comment

martinjaeger left a comment

Choose a reason for hiding this comment

raul-klg commented May 7, 2021

alexanderwachter commented Dec 31, 2020 •

edited

Loading

erwango May 7, 2021 •

edited

Loading

KozhinovAlexander commented Mar 20, 2021 •

edited

Loading

raul-klg commented Mar 31, 2021 •

edited

Loading

alexanderwachter commented May 6, 2021 •

edited

Loading