RE: [PATCH 2/3] usb: dwc3: host: Do not use dma_set_coherent_mask

From: Sriram Dash
Date: Thu Oct 27 2016 - 01:43:57 EST


>From: Arnd Bergmann [mailto:arnd@xxxxxxxx]
>On Tuesday, October 25, 2016 4:26:27 PM CEST Sriram Dash wrote:
>> Do not require dma_set_coherent_mask for hcd
>>
>> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
>
>Aside from the comments I had for patch 3, you are doing two different things
>here:
>
>> diff --git a/drivers/usb/dwc3/dwc3-st.c b/drivers/usb/dwc3/dwc3-st.c
>> index 89a2f71..4d7439c 100644
>> --- a/drivers/usb/dwc3/dwc3-st.c
>> +++ b/drivers/usb/dwc3/dwc3-st.c
>> @@ -218,7 +218,6 @@ static int st_dwc3_probe(struct platform_device *pdev)
>> if (IS_ERR(regmap))
>> return PTR_ERR(regmap);
>>
>> - dma_set_coherent_mask(dev, dev->coherent_dma_mask);
>> dwc3_data->dev = dev;
>> dwc3_data->regmap = regmap;
>>
>
>This one was setting the mask for the device itself (incorrectly), so it can be
>removed along with the dma_coerce_mask_and_coherent() call in
>dwc3_exynos_probe, or as a separate patch. It is an independent cleanup.
>
>> --- a/drivers/usb/chipidea/core.c
>> +++ b/drivers/usb/chipidea/core.c
>> @@ -833,9 +833,6 @@ struct platform_device *ci_hdrc_add_device(struct
>> device *dev,
>>
>> - dma_set_coherent_mask(&pdev->dev, dev->coherent_dma_mask);
>>
>> ret = platform_device_add_resources(pdev, res, nres);
>> if (ret)
>
>> --- a/drivers/usb/dwc3/core.c
>> +++ b/drivers/usb/dwc3/core.c
>> @@ -1059,12 +1059,6 @@ static int dwc3_probe(struct platform_device
>> *pdev)
>>
>> spin_lock_init(&dwc->lock);
>>
>> - dma_set_coherent_mask(dev, dev->parent-
>>coherent_dma_mask);
>
>> --- a/drivers/usb/dwc3/host.c
>> +++ b/drivers/usb/dwc3/host.c
>> @@ -72,11 +72,7 @@ int dwc3_host_init(struct dwc3 *dwc)
>> return -ENOMEM;
>> }
>>
>> - dma_set_coherent_mask(&xhci->dev, dwc->dev->coherent_dma_mask);
>> -
>
>These three all set the mask for the *child* devices, as that is no longer needed
>after the change in patch 1/3.
>I'd suggest leaving those changes together with the rest of that patch.
>
>However, it's probably better to split up that patch along the boundaries of the
>drivers, starting with the USB core:
>

Ok. Will do the needful in the next version.

>1/4 usb: separate out sysdev pointer from usb_bus
>2/4 usb: chipidea: use bus->sysdev for DMA configuration
>3/4 usb: xhci: use bus->sysdev for DMA configuration
>4/4 usb: dwc3: use bus->sysdev for DMA configuration
>
> Arnd