Re: [PATCH] tty: serial: fsl_lpuart: use dmaengine_get_dma_device() for DMA mapping

[Frank Li <Frank.li@oss.nxp.com>]

On Wed, Sep 16, 2026 at 11:27:30PM +0800, Peng Fan (OSS) wrote:
> From: Peng Fan <peng.fan@nxp.com>
>
> Replace all chan->device->dev references with
> dmaengine_get_dma_device(chan) so that DMA mapping goes through the
> correct device when the DMA controller has per-channel IOMMU domains
> (chan_dma_dev=true). Without this, DMA map/unmap/sync bypasses the
> IOMMU context configured by the eDMA driver.
>
> Signed-off-by: Peng Fan <peng.fan@nxp.com>
> ---

Reviewed-by: Frank Li <Frank.Li@nxp.com>

>  drivers/tty/serial/fsl_lpuart.c | 18 +++++++++---------
>  1 file changed, 9 insertions(+), 9 deletions(-)
>
> diff --git a/drivers/tty/serial/fsl_lpuart.c b/drivers/tty/serial/fsl_lpuart.c
> index b7919c05f0fb8..c8575c9652035 100644
> --- a/drivers/tty/serial/fsl_lpuart.c
> +++ b/drivers/tty/serial/fsl_lpuart.c
> @@ -491,7 +491,7 @@ static void lpuart_dma_tx(struct lpuart_port *sport)
>  	sport->dma_tx_nents = kfifo_dma_out_prepare(&tport->xmit_fifo, sgl,
>  			ARRAY_SIZE(sport->tx_sgl), sport->dma_tx_bytes);
>
> -	ret = dma_map_sg(chan->device->dev, sgl, sport->dma_tx_nents,
> +	ret = dma_map_sg(dmaengine_get_dma_device(chan), sgl, sport->dma_tx_nents,
>  			 DMA_TO_DEVICE);
>  	if (!ret) {
>  		dev_err(dev, "DMA mapping error for TX.\n");
> @@ -502,7 +502,7 @@ static void lpuart_dma_tx(struct lpuart_port *sport)
>  					ret, DMA_MEM_TO_DEV,
>  					DMA_PREP_INTERRUPT);
>  	if (!sport->dma_tx_desc) {
> -		dma_unmap_sg(chan->device->dev, sgl, sport->dma_tx_nents,
> +		dma_unmap_sg(dmaengine_get_dma_device(chan), sgl, sport->dma_tx_nents,
>  			      DMA_TO_DEVICE);
>  		dev_err(dev, "Cannot prepare TX slave DMA!\n");
>  		return;
> @@ -535,7 +535,7 @@ static void lpuart_dma_tx_complete(void *arg)
>  		return;
>  	}
>
> -	dma_unmap_sg(chan->device->dev, sgl, sport->dma_tx_nents,
> +	dma_unmap_sg(dmaengine_get_dma_device(chan), sgl, sport->dma_tx_nents,
>  		     DMA_TO_DEVICE);
>
>  	uart_xmit_advance(&sport->port, sport->dma_tx_bytes);
> @@ -606,7 +606,7 @@ static void lpuart_flush_buffer(struct uart_port *port)
>
>  	if (sport->lpuart_dma_tx_use) {
>  		if (sport->dma_tx_in_progress) {
> -			dma_unmap_sg(chan->device->dev, &sport->tx_sgl[0],
> +			dma_unmap_sg(dmaengine_get_dma_device(chan), &sport->tx_sgl[0],
>  				sport->dma_tx_nents, DMA_TO_DEVICE);
>  			sport->dma_tx_in_progress = false;
>  		}
> @@ -1176,7 +1176,7 @@ static void lpuart_copy_rx_to_tty(struct lpuart_port *sport)
>  	}
>
>  	/* CPU claims ownership of RX DMA buffer */
> -	dma_sync_sg_for_cpu(chan->device->dev, &sport->rx_sgl, 1,
> +	dma_sync_sg_for_cpu(dmaengine_get_dma_device(chan), &sport->rx_sgl, 1,
>  			    DMA_FROM_DEVICE);
>
>  	/*
> @@ -1237,7 +1237,7 @@ static void lpuart_copy_rx_to_tty(struct lpuart_port *sport)
>  	sport->last_residue = state.residue;
>
>  exit:
> -	dma_sync_sg_for_device(chan->device->dev, &sport->rx_sgl, 1,
> +	dma_sync_sg_for_device(dmaengine_get_dma_device(chan), &sport->rx_sgl, 1,
>  			       DMA_FROM_DEVICE);
>
>  	uart_port_unlock_irqrestore(&sport->port, flags);
> @@ -1374,7 +1374,7 @@ static inline int lpuart_start_rx_dma(struct lpuart_port *sport)
>  		return -ENOMEM;
>
>  	sg_init_one(&sport->rx_sgl, ring->buf, sport->rx_dma_rng_buf_len);
> -	nent = dma_map_sg(chan->device->dev, &sport->rx_sgl, 1,
> +	nent = dma_map_sg(dmaengine_get_dma_device(chan), &sport->rx_sgl, 1,
>  			  DMA_FROM_DEVICE);
>
>  	if (!nent) {
> @@ -1430,7 +1430,7 @@ static inline int lpuart_start_rx_dma(struct lpuart_port *sport)
>  	return 0;
>
>  err_unmap_sg:
> -	dma_unmap_sg(chan->device->dev, &sport->rx_sgl, 1, DMA_FROM_DEVICE);
> +	dma_unmap_sg(dmaengine_get_dma_device(chan), &sport->rx_sgl, 1, DMA_FROM_DEVICE);
>  err_free_buf:
>  	kfree(ring->buf);
>  	ring->buf = NULL;
> @@ -1447,7 +1447,7 @@ static void lpuart_dma_rx_free(struct uart_port *port)
>  	if (!sport->dma_idle_int)
>  		timer_delete_sync(&sport->lpuart_timer);
>
> -	dma_unmap_sg(chan->device->dev, &sport->rx_sgl, 1, DMA_FROM_DEVICE);
> +	dma_unmap_sg(dmaengine_get_dma_device(chan), &sport->rx_sgl, 1, DMA_FROM_DEVICE);
>  	kfree(sport->rx_ring.buf);
>  	sport->rx_ring.tail = 0;
>  	sport->rx_ring.head = 0;
>
> ---
> base-commit: e6e35979777d646fe3c7c94dca7dd32fb25d45f4
> change-id: 20260916-lpuart-dma-9923b7b2e1da
>
> Best regards,
> --
> Peng Fan <peng.fan@nxp.com>
>
>