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

From: Peter Ujfalusi
Date: Tue May 03 2016 - 16:00:57 EST


On 05/03/16 11:46, Ulf Hansson wrote:
> On 29 April 2016 at 15:06, Peter Ujfalusi <peter.ujfalusi@xxxxxx> wrote:
>> With the new dma_request_chan() the client driver does not need to look for
>> the DMA resource and it does not need to pass filter_fn anymore.
>> By switching to the new API the driver can now support deferred probing
>> against DMA.
>>
>> Signed-off-by: Peter Ujfalusi <peter.ujfalusi@xxxxxx>
>> CC: Ulf Hansson <ulf.hansson@xxxxxxxxxx>
>> CC: Jarkko Nikula <jarkko.nikula@xxxxxxxxxx>
>> ---
>> drivers/mmc/host/omap.c | 45 ++++++++++++++++++++++-----------------------
>> 1 file changed, 22 insertions(+), 23 deletions(-)
>>
>> diff --git a/drivers/mmc/host/omap.c b/drivers/mmc/host/omap.c
>> index b9958a123594..a8d9228657b2 100644
>> --- a/drivers/mmc/host/omap.c
>> +++ b/drivers/mmc/host/omap.c
>> @@ -23,7 +23,6 @@
>> #include <linux/spinlock.h>
>> #include <linux/timer.h>
>> #include <linux/of.h>
>> -#include <linux/omap-dma.h>
>> #include <linux/mmc/host.h>
>> #include <linux/mmc/card.h>
>> #include <linux/mmc/mmc.h>
>> @@ -1321,8 +1320,6 @@ static int mmc_omap_probe(struct platform_device *pdev)
>> struct omap_mmc_platform_data *pdata = pdev->dev.platform_data;
>> struct mmc_omap_host *host = NULL;
>> struct resource *res;
>> - dma_cap_mask_t mask;
>> - unsigned sig = 0;
>> int i, ret = 0;
>> int irq;
>>
>> @@ -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.

>
>> + dev_warn(host->dev, "TX DMA channel request failed\n");
>> + }
>> +
>> + host->dma_rx = dma_request_chan(&pdev->dev, "rx");
>> + if (IS_ERR(host->dma_rx)) {
>> + ret = PTR_ERR(host->dma_rx);
>> + if (ret == -EPROBE_DEFER) {
>> + dma_release_channel(host->dma_tx);
>
> host->dma_tx can be NULL here, so this isn't safe.

Oh, true.

>
>> + clk_put(host->fclk);
>> + goto err_free_iclk;
>> + }
>> +
>> + host->dma_rx = NULL;
>
> Instead of setting this to NULL, let's keep its value and later check
> it with IS_ERR() when needed.
>
>> + dev_warn(host->dev, "RX DMA channel request failed\n");
>> + }
>>
>> ret = request_irq(host->irq, mmc_omap_irq, 0, DRIVER_NAME, host);
>> if (ret)
>> --
>> 2.8.1
>>
>
> Kind regards
> Uffe
>


--
PÃter