[DEV] STM32G4xx support

[MasterSpoon]

Not sure if this is the correct place to ask but it is about STM32G4xx support.

Is there a current ETA on when the STM32G491xx and STM32G4A1xx range will be supported by ChibiOS ?

They are the higher capacity variants of the STM32G431xx's which are already supported and sit between the G431 and G471 in terms of features.
https://www.st.com/en/microcontrollers- ... 2g4x1.html

I can see the header files are already available from ST
https://github.com/STMicroelectronics/S ... xx/Include

[Giovanni]

Hi,

Problem is availability of HW mainly.

Giovanni

[preisi]

Hello,
recently I've been trying to get FDCAN to work with a STM32G474 but so far, I didn't succeed. Thus, I was wondering what the current state of FDCAN for STM32G4xx actually is. Reading through this thread was not very conclusive in my opinion (at some point people only wrote about STM32Hxxx support?).
Therefore my question: Is FDCAN already working with STM32G4xx on upstream? If that's the case, is there any working example/documentation?
Best regards,
preisi

[Giovanni]

Hi,

FDCAN is not yet functional, several people have proposed patches but none has yet been integrated or tested.

Giovanni

[preisi]

Hello,
thanks for your response. At least now I know why it didn't work
Is at least 'normal' CAN functional at the moment?
Best regards,
preisi

[Giovanni]

Hi,

The normal CAN driver works but that peripheral is not present on the G4 family.

Giovanni

[Guillaume227]

I am wondering about ADC5 support : I see earlier in this thread that support was not added at the same time as ADC 1...4. What's different about ADC5 that we can't mimic what's been done for the other ADC in its group, namely ADC3 and ADC4?

[Giovanni]

I am wondering about ADC5 support : I see earlier in this thread that support was not added at the same time as ADC 1...4. What's different about ADC5 that we can't mimic what's been done for the other ADC in its group, namely ADC3 and ADC4?

There is nothing preventing it except it could not be used in dual mode because there is no ADC6.

Giovanni

[Apache]

I just had a look at this diff, and it looks like I actually already made all these changes in my branch, except for adding the call to fdcan_clock_start. Do you really need to add that? It works just fine on an H7 (same IP from Bosch) exactly as it is on my fork (https://github.com/mck1117/chibios/tree/h7-fdcan).

I base my work on yours, but it still didn't wor for my mcu (STM32G431). I activated the asserts and realized that the clock was not started after configuring the peripheral. It don't understand why it worked on your side

I just added the clock start to your fixes and it started working right away. So the diff I sent is just your diff with that enhancement

[Apache]

Hello everyone,
I noticed a new problem in the implementation of the low level driver: it doesn't handle extended IDs. I quick fixed it so I could use it but I didn't code a clean fix yet. If anyone has the same problem, the fix is to set the bit 30 of the TX buffer, which allows the peripheral to send extended IDs.

At some point I'll be back with a real patch and we'll be able to discuss it.

Greets