Problem configuring DMA on STM32F130rb trying to read a temp sensor

[nowave7]

Hi all,

I'm trying to write a small program that will read the values from the temperature/humidity sensor. I will try to explain in detail what I have at the moment, but bear with me, I'm still a beginner in bare metal development.

I am using an STM32F103RB board, DHT22 temp/hum sensor, and FreeRTOS as middleware. The environment is STM32CubeIDE (eclipse based, arm gcc compiler, st-flash) with HAL driver and CMSIS BSP.

I have configured the TIM3_CH1/Trig on channel 6 and on pin PA6. Since the sensor is using a singel wire protocol, I am switching the mode of the GPIO from output to input and back in order to get the data in. The problem occurs when I call the HAL_TIM_IC_Start_DMA function. In it, the correct channel will be chosen, and the corresponding callbacks should be set. And that is exactly where I get the a hard fault interrupt. in particular at this specific line:

htim->hdma[TIM_DMA_ID_CC3]->XferHalfCpltCallback = TIM_DMACaptureHalfCplt;

in file stm32f1xx_hal_tim.c, line 2106.

I've added the handler in the HardFault_Handler, that gets the status of the registers, but, as a novice, that doesn't mean that much to me.

So, at the moment, I am stuck at this point.

I have included the entire project as an attachment, hopefully, it is not too much to ask, someone could have a look at it.

Thank you all in advance!

Vladimir

[waclawek.jan]

> I get the a hard fault interrupt. in particular at this specific line:

> htim->hdma[TIM_DMA_ID_CC3]->XferHalfCpltCallback = TIM_DMACaptureHalfCplt;

Single-step in disasm. Does the hardfault happen at the same instruction? Which one, and what's in its vicinity (best in mixed C/asm view)?

> I've added the handler in the HardFault_Handler, that gets the status of the registers, but, as a novice, that doesn't mean that much to me.

OK so post the info together with that above.

I don't Cube/HAL.

JW

[nowave7]

Thank you kindly for the reply Jan.

After running the debugger in mixed mode with instruction stepping mode on, I can see that the interrupt always occurs on the same instruction:

2125      htim->hdma[TIM_DMA_ID_CC3]->XferHalfCpltCallback = TIM_DMACaptureHalfCplt;

080015e8:  ldr   r3, [r7, #12]

Oddly enough, the same instruction was executed for the previous C statment:

2124      htim->hdma[TIM_DMA_ID_CC3]->XferCpltCallback = TIM_DMACaptureCplt;

080015e0:  ldr   r3, [r7, #12]

and there was no problem at all.

I've attached an image with the listing of both C and asm code, and the contents of the registers right before executing the instruction in question.

[nowave7]

After some digging around, the instruction that causes the hard fault (ldr   r3, [r7, #12]) is nothing special at all, seems to be quite common throughout the code. But after assigning the first callback htim->hdma[TIM_DMA_ID_CC3]->XferCpltCallback = TIM_DMACaptureCplt; it seems to start causing issues. Since I don't really need to know when the data is half way received, I commented out assigning the first callback, just to see what happens. Assigning the second callback (TIM_DMACaptureCplt) succeeds, only to hard fault on the next such instruction.