Re: [PATCH] mmc: omap: Use dma_request_chan() for requesting DMA channel

[Peter Ujfalusi <peter.ujfalusi@ti.com>]

On 05/03/2016 11:00 PM, Peter Ujfalusi wrote:
>>> @@ -1382,29 +1379,31 @@ static int mmc_omap_probe(struct platform_device *pdev)
>>>                 goto err_free_iclk;
>>>         }
>>>
>>> -       dma_cap_zero(mask);
>>> -       dma_cap_set(DMA_SLAVE, mask);
>>> -
>>>         host->dma_tx_burst = -1;
>>>         host->dma_rx_burst = -1;
>>>
>>> -       res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "tx");
>>> -       if (res)
>>> -               sig = res->start;
>>> -       host->dma_tx = dma_request_slave_channel_compat(mask,
>>> -                               omap_dma_filter_fn, &sig, &pdev->dev, "tx");
>>> -       if (!host->dma_tx)
>>> -               dev_warn(host->dev, "unable to obtain TX DMA engine channel %u\n",
>>> -                       sig);
>>> -
>>> -       res = platform_get_resource_byname(pdev, IORESOURCE_DMA, "rx");
>>> -       if (res)
>>> -               sig = res->start;
>>> -       host->dma_rx = dma_request_slave_channel_compat(mask,
>>> -                               omap_dma_filter_fn, &sig, &pdev->dev, "rx");
>>> -       if (!host->dma_rx)
>>> -               dev_warn(host->dev, "unable to obtain RX DMA engine channel %u\n",
>>> -                       sig);
>>> +       host->dma_tx = dma_request_chan(&pdev->dev, "tx");
>>> +       if (IS_ERR(host->dma_tx)) {
>>> +               ret = PTR_ERR(host->dma_tx);
>>> +               if (ret == -EPROBE_DEFER)
>>> +                       goto err_free_iclk;
> 
> and one clk_put(host->fclk) is missing from here.
> 
>>> +
>>> +               host->dma_tx = NULL;
>>
>> Instead of setting this to NULL, let's keep its value and later check
>> it with IS_ERR() when needed.
> 
> I thought about it, but decided against without need to have changes in other
> places in the driver.

This sentence does not make too much sense ;) Let's try it again:
So, I have considered to use PTR_ERR in the driver for the DMA channel, but I
wanted to avoid too broad changes in the driver. Keeping the ERR_PTR would
bring no runtime benefit since we only care if we have channel or not.

-- 
Péter