Перед считыванием данных я проверяю статус (CMD13) - если карта в TRAN, то буду считывать:

else if(cs.current_state == CARD_STATE_TRAN)
{
SDIO->DCTRL = 0;
SDIO->DLEN = sdio_p_buf_size;
DMA2_Stream6->CR = 0
| (4 << DMA_SCR_CHSEL)
| (1 << DMA_SCR_MINC)
| (2 << DMA_SCR_MSIZE)
| (2 << DMA_SCR_PSIZE)
| (1 << DMA_SCR_PBURST)
| (1 << DMA_SCR_PFCTRL)
| (0 << DMA_SCR_EN);
DMA2_Stream6->FCR = 0
| (1 << DMA_SFCR_DMDIS)
| (1 << DMA_SFCR_FTH);
DMA2->HIFCR = (0x3D << 16);
DMA2_Stream6->PAR = (DWORD)&SDIO->FIFO;
DMA2_Stream6->M0AR = (DWORD)sdio_p_buf;
DMA2_Stream6->NDTR = sdio_p_buf_size / sizeof(DWORD);
sdio_fsm = SDIO_FSM_READ;
SDIO_Send(CMD18_READ_MULTIPLE_BLOCK, (sdio_sect << ((card_type == CARD_SDHC)? 0 : 9)), EXPECT_SHORT_RESP);
}

когда CMD18 выполнится запускаю прием данных так:

DMA2_Stream6->CR = 0
| (4 << DMA_SCR_CHSEL)
| (1 << DMA_SCR_MINC)
| (2 << DMA_SCR_MSIZE)
| (2 << DMA_SCR_PSIZE)
| (1 << DMA_SCR_PBURST)
| (1 << DMA_SCR_PFCTRL)
| (1 << DMA_SCR_EN);
SDIO->DCTRL = 0
| (9 << SDIO_DCTRL_DBLOCKSIZE)
| (1 << SDIO_DCTRL_DTDIR)
| (1 << SDIO_DCTRL_DMAEN)
| (1 << SDIO_DCTRL_DTEN);

по приходу прерывания DATAEND отправляю "горшочек не вари" для CMD18:

if(status & (1 << SDIO_DATAEND))
{
SDIO->ICR = (1 << SDIO_DATAEND);
SDIO->DCTRL = 0;
sdio_fsm = SDIO_FSM_READ_STOP;
SDIO_Send(CMD12_STOP_TRANSMISSION, NO_ARGUMENTS, EXPECT_SHORT_RESP);
}

    BSP_SD_Init();
    char buffer[512];
    BSP_SD_ReadBlocks((uint32_t *)buffer, 0, 512, 1);

hdma_sdio.Instance = DMA2_Stream6;
hdma_sdio.Init.Channel = DMA_CHANNEL_4;
hdma_sdio.Init.Direction = DMA_MEMORY_TO_PERIPH;//DMA_PERIPH_TO_MEMORY
hdma_sdio.Init.PeriphInc = DMA_PINC_DISABLE;
hdma_sdio.Init.MemInc = DMA_MINC_ENABLE;
hdma_sdio.Init.PeriphDataAlignment = DMA_PDATAALIGN_WORD;
hdma_sdio.Init.MemDataAlignment = DMA_MDATAALIGN_WORD;
hdma_sdio.Init.Mode = DMA_PFCTRL;
hdma_sdio.Init.Priority = DMA_PRIORITY_LOW; // Единственное что не имеет значения
hdma_sdio.Init.FIFOMode = DMA_FIFOMODE_ENABLE;
hdma_sdio.Init.FIFOThreshold = DMA_FIFO_THRESHOLD_FULL;
hdma_sdio.Init.MemBurst = DMA_MBURST_INC4;
hdma_sdio.Init.PeriphBurst = DMA_PBURST_INC4;
HAL_DMA_Init(&hdma_sdio);

__HAL_LINKDMA(hsd,hdmarx,hdma_sdio);
__HAL_LINKDMA(hsd,hdmatx,hdma_sdio);

static BOOL WriteBlock (U32 bl, U8 *buf, U32 cnt) {
  /* Write a cnt number of 512 byte blocks to Flash Card. */
  U32 i;

  SDIO->DLEN    = cnt * 512;
  SDIO->DTIMER  = cnt * DATA_WR_TOUT_VALUE;
  SDIO->DCTRL   = SDIO_DCTRL_DBLOCKSIZE_3 | SDIO_DCTRL_DBLOCKSIZE_0 |
                  SDIO_DCTRL_DMAEN        | SDIO_DCTRL_DTEN    ;

  for (i = DMA_TOUT; i; i--) {
    if (DMA2->LISR & DMA_LISR_TEIF3) {
      break;
    }
    
    if (DMA2->LISR & DMA_LISR_TCIF3) {
      if ((SDIO->STA & (SDIO_STA_DBCKEND|SDIO_STA_DATAEND)) == (SDIO_STA_DBCKEND|SDIO_STA_DATAEND)) {
        /* Data transfer finished. */
        return (__TRUE);
      }
    }
  }
  /* DMA Transfer timeout. */
  return (__FALSE);
}