Re: [PATCH net] macmace: Set platform device coherent_dma_mask

[Finn Thain]

On Fri, 11 May 2018, Michael Schmitz wrote:

> 
> I wasn't proposing to follow platform_device_register() with the mask 
> assignment,

I see.

> but rather to use the same strategy from the Coldfire FEC patch 
> (f61e64310b75): set pdev.dev.coherent_dma_mask and pdev.dev.dma_mask 
> _before_ calling platform_device_register().
> 

That patch is attempting to solve the same problem using a different 
approach, but I'd prefer to keep using platform_device_register_simple().

> 
> If you want to avoid the overhead of defining the macmace platform 
> device and using platform_device_register() ... seeing as you would not 
> be defining just the DMA mask and nothing else, that's probably the 
> least effort for the MACE and SONIC drivers.
> 

I don't mind the effort. I want to avoid all the boilerplate.

> 
> You would have to be careful not to overwrite a pdev->dev.dma_mask and 
> pdev->dev.dma_coherent_mask that might have been set in a platform 
> device passed via platform_device_register here. Coldfire is the only 
> m68k platform currently using that, but there might be others in future.
> 

That Coldfire patch could be reverted if this is a better solution.

> ... But I don't think there are smaller DMA masks used by m68k drivers 
> that use the platform device mechanism at present. I've only looked at 
> arch/m68k though.

So we're back at the same problem that Geert's suggestion also raised: how 
to identify potentially affected platform devices and drivers?

Maybe we can take a leaf out of Christoph's book, and leave a noisy 
WARNING splat in the log.

void arch_setup_pdev_archdata(struct platform_device *pdev)
{
        WARN_ON_ONCE(pdev->dev.coherent_dma_mask != DMA_MASK_NONE ||
                     pdev->dev.dma_mask != NULL);
        pdev->dev.coherent_dma_mask = DMA_BIT_MASK(32);
        pdev->dev.dma_mask = &pdev->dev.coherent_dma_mask;
}

--