[PATCH 06/10] soc/qbman: Add ARM equivalent for flush_dcache_range()

[Roy Pledge]

On 2/1/2017 8:03 AM, Robin Murphy wrote:
>
>> My understanding is that the dma_declare_coherent_memory() API sets up a
>> region of memory that will be managed by dma_alloc_coherent() and
>> friends.  This is useful if the driver needs to manage a region of on
>> device memory but that isn't what this specific region is used for.
> It's a bit more general than that - dma_alloc_coherent() can essentially
> be considered "give me some memory for this device to use". We already
> have use-cases where such buffers are only ever accessed by the device
> (e.g. some display controllers, and future NVMe devices), hence
> DMA_ATTR_NO_KERNEL_MAPPING on ARM to save the vmalloc space.
>
> A DMA allocation also inherently guarantees appropriate alignment,
> regardless of whether you're using a per-device reservation or just
> regular CMA, and will also zero the underlying memory (and for a
> non-coherent device perform whatever cache maintenance is necessary, if
> the clearing isn't already done via a non-cacheable mapping).
>
> All you need to do in the driver is allocate your buffer and hand the
> resulting address off to the device at probe (after optionally checking
> for a reservation in DT and declaring it), then free it at remove, which
> also ends up far more self-documenting (IMO) than a bunch of open-coded
> remapping and #ifdef'ed architecture-private cache shenanigans.
I found some time to work on this but the dma_declare_coherent_memory()
API does not allow me to pass normal memory for the device to use.

Depending on which flag is used I get one of two behaviors. If I pass
DMA_MEMORY_MAP the path attempts to do a memremap() which fails with the
message "memremap attempted on ram". If I use DMA_MEMORY_IO then it
attempts to use ioremap() on the memory which fails for a similar reason
(Not allowed to map ram).  Using dma_alloc_coherent() fails in our basic
use case because it is unable to obtain 8 megabytes of RAM.

Why don't I simplify this and use dma_sync_single_for_device() to flush
the cache? This seems to me to be portable and the simplest solution to
this problem. The device reports non-coherent when queried using 
is_device_dma_coherent() and my very quick walk through the code seems
to indicate that the call will do the right thing.

Roy

Roy