Re: [PATCH] tty: serial: 8250_omap: fix tx with dma

[Jiri Slaby <jirislaby@kernel.org>]

On 30. 04. 25, 18:37, Mans Rullgard wrote:
> Commit 1788cf6a91d9 ("tty: serial: switch from circ_buf to kfifo")
> introduced two errors in the TX DMA handling for 8250_omap.
> 
> Firstly, kfifo_dma_out_prepare_mapped() needs a scatterlist with two
> entries whereas only one is provided.  The same error was fixed for
> 8250_dma in 59449c9dbdaa ("tty: serial: 8250_dma: use sgl with 2 nents
> to take care of buffer wrap").

It's not an error. This is how it used to work since ever. Providing two 
is an optimization, right?

> Secondly, when the OMAP_DMA_TX_KICK flag is set, one byte is pulled from
> the kfifo and emitted directly in order to start the DMA.

> This is done without updating DMA tx_size

Ah, that's an error, of course.

> which leads to uart_xmit_advance() called
> in the DMA complete callback advancing the kfifo by one too much.
> 
> In practice, transmitting N bytes has been seen to result in the last
> N-1 bytes being sent repeatedly.
> 
> This change fixes both problems.

I am not sure you want to mix fixups with features.

> Fixes: 1788cf6a91d9 ("tty: serial: switch from circ_buf to kfifo")
> Signed-off-by: Mans Rullgard <mans@mansr.com>
...
> @@ -1248,8 +1247,10 @@ static int omap_8250_tx_dma(struct uart_8250_port *p)
>   err:
>   	dma->tx_err = 1;
>   out_skip:
> -	if (skip_byte >= 0)
> +	if (skip_byte >= 0) {
>   		serial_out(p, UART_TX, skip_byte);
> +		p->port.icount.tx++;

This is unrelated (but correct) too.

thanks,
-- 
js
suse labs