On 2019-11-13 8:41 pm, Florian Fainelli wrote:

On 11/13/19 12:34 PM, Robin Murphy wrote:

quoted

On 13/11/2019 4:13 pm, Nicolas Saenz Julienne wrote:

quoted

Using a mask to represent bus DMA constraints has a set of limitations.
The biggest one being it can only hold a power of two (minus one). The
DMA mapping code is already aware of this and treats dev->bus_dma_mask
as a limit. This quirk is already used by some architectures although
still rare.

With the introduction of the Raspberry Pi 4 we've found a new contender
for the use of bus DMA limits, as its PCIe bus can only address the
lower 3GB of memory (of a total of 4GB). This is impossible to represent
with a mask. To make things worse the device-tree code rounds non power
of two bus DMA limits to the next power of two, which is unacceptable in
this case.

In the light of this, rename dev->bus_dma_mask to dev->bus_dma_limit all
over the tree and treat it as such. Note that dev->bus_dma_limit is
meant to contain the higher accesible DMA address.

Neat, you win a "why didn't I do it that way in the first place?" :)

Looking at it without all the history of previous attempts, this looks
entirely reasonable, and definitely a step in the right direction.

And while you are changing those, would it make sense to not only rename
the structure member but introduce a getter and setter in order to ease
future work where this would no longer be a scalar?

I doubt it - once we get as a far as supporting multiple DMA ranges, 
there will be a whole load of infrastructure churn anyway if only to 
replace dma_pfn_offset, and I'm not sure a simple get/set paradigm would 
even be viable, so it's probably better to save that until clearly 
necessary.

Robin.