On 17-11-21, 14:20, Doug Anderson wrote:
> Hi,
>
> On Wed, Nov 17, 2021 at 5:31 AM Vinod Koul <vkoul@xxxxxxxxxx> wrote:
> >
> > @@ -346,17 +346,20 @@ spi_gsi_callback_result(void *cb, const struct dmaengine_result *result)
> > {
> > struct spi_master *spi = cb;
> >
> > + spi->cur_msg->status = -EIO;
> > if (result->result != DMA_TRANS_NOERROR) {
> > dev_err(&spi->dev, "DMA txn failed: %d\n", result->result);
> > return;
> > }
>
> Don't you want to call spi_finalize_current_transfer() in the case of
> a DMA error? Otherwise I think you're still going to wait for a
> timeout? ...and then when you get the timeout then spi_transfer_wait()
> will return -ETIMEDOUT and that will overwrite your -EIO, won't it?
Yes missed that and this reply :(
Fixed now and posting v2
--
~Vinod
