drivers: adc: stm32: use correct device in dma callback

[bwitherspoon]

The device passed to the DMA callback is the DMA device. Instead use the ADC device available in the private data.

[gautierg-st]

@heinwessels, you worked on the DMA part for the STM32 ADC. Would you mind taking a look?

[bwitherspoon]

For a little background, I discovered this issue while adding ADC DMA support for the nucleo_u575zi_q board which uses the GPDMA and this causes a fault. I checked the other STM32 DMA drivers including the new BDMA and the DMA device is passed to the callback.

I will be submitting other fixes for the GPDMA driver which includes adding the nucleo_u575zi_q board to the ADC DMA tests if that is helpful to evaluate this change. But this fix seems like it could go ahead and be submitted.

[heinwessels]

Seems I did get it wrong in the initial implementation. Nice catch!

[gautierg-st]

For a little background, I discovered this issue while adding ADC DMA support for the nucleo_u575zi_q board which uses the GPDMA and this causes a fault. I checked the other STM32 DMA drivers including the new BDMA and the DMA device is passed to the callback.

I will be submitting other fixes for the GPDMA driver which includes adding the nucleo_u575zi_q board to the ADC DMA tests if that is helpful to evaluate this change. But this fix seems like it could go ahead and be submitted.

All contributions are welcome. Concerning the DMA on ADC, ideally its support should be added for all STM32 series. That's in my to-do list of ADC driver improvements, but not in the immediate future, so if you have something working, even if only for U5, you're welcome.