USB CDC Receive for the STM32F103C8 using CubeMX?

Question

Posted on September 06, 2017 at 02:55

How should I implement USB CDC Receive for the STM32F103C8 using CubeMX? I could not find any examples. 
 I have a partially working implementation of a USB CDC for this IC, but I have a few occasional issues with missed data. 
 First, I don't understand the
 CDC_Receive_FS(...) function. 
 Do I process/copy the buffer on line 5 or line 10?Why does it need to SetRxBuffer? It shouldn't change, should it? Isn't it always stored intoUserRxBufferFS?Or is this function passing a pointer to some other buffer internal to the USB driver, and UserRxBufferFS is never used?What does USB_CDC_RecivePacket do? Since it does LL stuff, why wouldn't it be called by the function that calls CDC_Receive_FS?My understanding is that the max packet length for full-speed USB is 512 bytes. Do I need to changeAPP_RX_DATA_SIZE to be 512?Are there still concerns with read/write locking? (Seehttps://community.st.com/thread/10392).Is there a way to cancel the previous USB transmit, or append data to its buffer? 
 static int8_t CDC_Receive_FS (uint8_t* Buf, uint32_t *Len)
{
 /* USER CODE BEGIN 6 */
 
 /**** Do I process the buffer here? ****/
 /* Can the next line be removed, unless the RX buffer needs to change? */
 USBD_CDC_SetRxBuffer(&hUsbDeviceFS, &Buf[0]); /*<-- Change buffer for next RX? */
 USBD_CDC_ReceivePacket(&hUsbDeviceFS);
 
 /**** Or do I process it here? ****/
 return (USBD_OK);
 /* USER CODE END 6 */ 
}�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�? 
 I'm currently detecting if the USB CDC is connected by seeing if the previous transmission succeeded in a timely fashion. I'm using FreeRTOS. Is this proper, or is there a better way? Since I don't know how to cancel the previous transmit, I don't want to overwrite the buffer while it may still be transmitting. 
 static void usbTx(char *msg) {
 static TickType_t lastXmit = 0;
 /* If still transmitting, and it has been more than 200 ms, give up */
 while(CDC_Busy() && (xTaskGetTickCount()-lastXmit < (200/portTICK_PERIOD_MS ))) {
 osDelay(1);
 }
 // Last transmission failed, so lets skip this one.
 if(CDC_Busy())
 return;
 size_t len = strlen(msg);
 uint8_t result = CDC_Transmit_FS((uint8_t*)msg,len);
 if(result == USBD_OK)
 lastXmit = xTaskGetTickCount();
}�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�? 
 #freertos-stm32cube #stm32f103 #cdc #usb

Nathan Conrad · Answer

Posted on September 10, 2017 at 05:32 (I'm replying to my own post, though I'd still appreciate answers to the questions I posed.)  I found that I had a memory-management bug in my code while transmitting. The issue was that after calling the transmit function, the TX buffer may not change until the transmission is complete. I've not done thorough testing, but my CDC implementation seems to be reliable now.  To solve my bug, I added a memcpy(...) to the TX function:  // Use 65, just to be safe. I dislike buffer overruns.#define APP_TX_DATA_SIZE 65uint8_t CDC_Transmit_FS(uint8_t* Buf, uint16_t Len){ uint8_t result = USBD_OK; /* USER CODE BEGIN 7 */ if(hUsbDeviceFS.dev_state != USBD_STATE_CONFIGURED)return USBD_FAIL;USBD_CDC_HandleTypeDef *hcdc = (USBD_CDC_HandleTypeDef*)hUsbDeviceFS.pClassData;if (hcdc->TxState != 0){return USBD_BUSY;}// I've read that lengths of 64 can confuse Windows, as it interprets it as a reset.configASSERT(Len > 0 && Len < 63);memcpy(UserTxBufferFS, Buf, Len);USBD_CDC_SetTxBuffer(&hUsbDeviceFS, UserTxBufferFS, Len);result = USBD_CDC_TransmitPacket(&hUsbDeviceFS); /* USER CODE END 7 */  return result;}�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?�?

Sign up

Login with SSO

Login to the community

Login with SSO