4 Dec
2014
4 Dec
'14
9:57 a.m.
On 12/03/14 16:44, Lars-Peter Clausen wrote:
On 12/03/2014 05:38 PM, Andrew Jackson wrote: [,,,[
- dw_i2s_dai = devm_kzalloc(&pdev->dev, sizeof(*dw_i2s_dai), GFP_KERNEL);
- if (!dw_i2s_dai) {
dev_err(&pdev->dev, "mem allocation failed for dai driver\n");All the memory alloc functions already print a error message.
I will remove the error message(s).
return -ENOMEM;}
- dw_i2s_dai->ops = &dw_i2s_dai_ops;
- dw_i2s_dai->suspend = dw_i2s_suspend;
- dw_i2s_dai->resume = dw_i2s_resume;
This seems to be separate from the devm_ioremap_resource() change.
It seemed reasonable to assign these known values to the dw_i2s_dai as soon as it had been allocated. Would you prefer this in a separate patch?
- res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
- if (!res) {
You don't actually have to check it devm_ioremap_resource does this for you.
Oh, thanks, paranoia was taking over.
dev_err(&pdev->dev, "no i2s resource defined\n");return -ENODEV;- }
- dev->i2s_base = devm_ioremap_resource(&pdev->dev, res);
- if (IS_ERR(dev->i2s_base)) {
dev_err(&pdev->dev, "ioremap fail for i2s_region\n");Same here devm_ioremap_resource() will already print a appropriate error message.
return PTR_ERR(dev->i2s_base);- }