This reverts commit 906b40b246b0acb54c4dc97e815cf734761c9820.
As stated by bindings "st,ahb-addr-masks" is optional.
The statement inserted by this commit makes this property
mandatory and prevents MDMA to be probed in case property not present.
Signed-off-by: Pierre-Yves MORDRET <[email protected]>
---
drivers/dma/stm32-mdma.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/dma/stm32-mdma.c b/drivers/dma/stm32-mdma.c
index 4e0eede..ac0301b 100644
--- a/drivers/dma/stm32-mdma.c
+++ b/drivers/dma/stm32-mdma.c
@@ -1578,11 +1578,9 @@ static int stm32_mdma_probe(struct platform_device *pdev)
dmadev->nr_channels = nr_channels;
dmadev->nr_requests = nr_requests;
- ret = device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
+ device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
dmadev->ahb_addr_masks,
count);
- if (ret)
- return ret;
dmadev->nr_ahb_addr_masks = count;
res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
--
2.7.4
On 25-03-19, 15:46, Pierre-Yves MORDRET wrote:
Please use the right subsystem name dmaengine, revert is not a subsystem
name!
> This reverts commit 906b40b246b0acb54c4dc97e815cf734761c9820.
This should use the cannonical form for commits commit-sha1: ("title....")
>
> As stated by bindings "st,ahb-addr-masks" is optional.
> The statement inserted by this commit makes this property
> mandatory and prevents MDMA to be probed in case property not present.
>
> Signed-off-by: Pierre-Yves MORDRET <[email protected]>
> ---
> drivers/dma/stm32-mdma.c | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/drivers/dma/stm32-mdma.c b/drivers/dma/stm32-mdma.c
> index 4e0eede..ac0301b 100644
> --- a/drivers/dma/stm32-mdma.c
> +++ b/drivers/dma/stm32-mdma.c
> @@ -1578,11 +1578,9 @@ static int stm32_mdma_probe(struct platform_device *pdev)
>
> dmadev->nr_channels = nr_channels;
> dmadev->nr_requests = nr_requests;
> - ret = device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
> + device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
> dmadev->ahb_addr_masks,
> count);
> - if (ret)
> - return ret;
> dmadev->nr_ahb_addr_masks = count;
so if st,ahb-addr-masks is not present what value does count contain?
>
> res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
> --
> 2.7.4
--
~Vinod
On 3/25/19 4:25 PM, Vinod Koul wrote:
> On 25-03-19, 15:46, Pierre-Yves MORDRET wrote:
>
> Please use the right subsystem name dmaengine, revert is not a subsystem
> name!
ok. sorry. I thought revert keyword was correct.
>
>> This reverts commit 906b40b246b0acb54c4dc97e815cf734761c9820.
>
> This should use the cannonical form for commits commit-sha1: ("title....")
ok.
>>
>> As stated by bindings "st,ahb-addr-masks" is optional.
>> The statement inserted by this commit makes this property
>> mandatory and prevents MDMA to be probed in case property not present.
>>
>> Signed-off-by: Pierre-Yves MORDRET <[email protected]>
>> ---
>> drivers/dma/stm32-mdma.c | 4 +---
>> 1 file changed, 1 insertion(+), 3 deletions(-)
>>
>> diff --git a/drivers/dma/stm32-mdma.c b/drivers/dma/stm32-mdma.c
>> index 4e0eede..ac0301b 100644
>> --- a/drivers/dma/stm32-mdma.c
>> +++ b/drivers/dma/stm32-mdma.c
>> @@ -1578,11 +1578,9 @@ static int stm32_mdma_probe(struct platform_device *pdev)
>>
>> dmadev->nr_channels = nr_channels;
>> dmadev->nr_requests = nr_requests;
>> - ret = device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
>> + device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
>> dmadev->ahb_addr_masks,
>> count);
>> - if (ret)
>> - return ret;
>> dmadev->nr_ahb_addr_masks = count;
>
> so if st,ahb-addr-masks is not present what value does count contain?
whether "st,ahb-addr-masks" present count will hold the number of mask.
otherwise sets to "0" by driver.
count = device_property_read_u32_array(&pdev->dev, "st,ahb-addr-masks",
NULL, 0);
if (count < 0)
count = 0;
>
>>
>> res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>> --
>> 2.7.4
>